VibePrevisBench
Coding agents write Blender Python that turns a script, photos, an opening frame or reference footage into a graybox 4D previs scene and render it. Rule checks score the exported scene against hidden facts, and a VLM judge checks that those facts are visible in the render.
242 tasks. The overall score is half Scene Check and half Screen Check on a 0 to 100 scale, with failed deliveries scored zero. The judge, Gemini 4 Argon, is also one of the 11 agents tested. Annotations are CC BY-NC-SA 4.0.
- Made by
- KlingAI Research (Kling-Keye)
- Released
- November 2026
- Kind
- CAD models
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Subject
- Computer and information sciences, Arts
- Language
- English
- Interaction
- Agentic
- Environment
- Blender
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Agents, Vision, Image, video and 3D generation
- Strongest model tested
- GPT-6 Astra #4 on the capability index
- Last model added
- September 2026
- Scores refreshed
- November 20th, 2026
GPT-6 Astra leads with 77.7 among models released by September 2026
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.
The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6.1 Terra
- and 11 more
- 1 GPT-6 Astra reasoning: max; 4-hour task limit 77.7
- 2 GPT-6.1 Sol reasoning: max; 4-hour task limit 76.0
- 3
Gemini 4 Argon reasoning: high; 4-hour task limit 75.0
- 4 Claude Opus 5.5 reasoning: high; 4-hour task limit 74.1
- 5 GPT-5.6 Sol reasoning: max; 4-hour task limit 70.2
- 6
DeepSeek V4.1 Flash reasoning: max; 4-hour task limit; model snapshot: 1017 69.6
- 7 Kimi K3 reasoning: xhigh; 4-hour task limit 67.5
- 8
Gemini 3.8 Flash reasoning: high; 4-hour task limit 64.7
- 9 GLM-5.3-Flash reasoning: high; 4-hour task limit 63.7
- 10
Doubao Seed 2.1 Pro reasoning: max; 4-hour task limit 61.3
- 11
Muse Spark 1.3 reasoning: xhigh; 4-hour task limit 61.1
Source: Official VibePrevisBench leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-6 Astra | 77.7 | reasoning: max; 4-hour task limit, Codex CLI | Sep 2026 | Official leaderboard |
| 2 | GPT-6.1 Sol | 76.0 | reasoning: max; 4-hour task limit, Codex CLI | Sep 2026 | Official leaderboard |
| 3 | | 75.0 | reasoning: high; 4-hour task limit, Claude Code | Sep 2026 | Official leaderboard |
| 4 | Claude Opus 5.5 | 74.1 | reasoning: high; 4-hour task limit, Claude Code | Sep 2026 | Official leaderboard |
| 5 | GPT-5.6 Sol | 70.2 | reasoning: max; 4-hour task limit, Codex CLI | Jul 2026 | Official leaderboard |
| 6 | | 69.6 | reasoning: max; 4-hour task limit; model snapshot: 1017, Claude Code | Sep 2026 | Official leaderboard |
| 7 | Kimi K3 | 67.5 | reasoning: xhigh; 4-hour task limit, Claude Code | Jul 2026 | Official leaderboard |
| 8 | | 64.7 | reasoning: high; 4-hour task limit, Claude Code | Sep 2026 | Official leaderboard |
| 9 | GLM-5.3-Flash | 63.7 | reasoning: high; 4-hour task limit, Claude Code | Aug 2026 | Official leaderboard |
| 10 | | 61.3 | reasoning: max; 4-hour task limit, Claude Code | Jun 2026 | Official leaderboard |
| 11 | | 61.1 | reasoning: xhigh; 4-hour task limit, Claude Code | Sep 2026 | Official leaderboard |
Rankings by part
We read 189 parts from the benchmark’s own leaderboard in 5 groups: modality, environment, dimension, category and length. Each ranks models on one part of the test. GPT-6 Astra leads 128 of them, and 6 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Text
11 models we track have a score here.
- 1 GPT-6 Astra 81.0
- 2 GPT-6.1 Sol 78.8
- 3 Claude Opus 5.5 ↑1 78.0
- 4
Gemini 4 Argon ↓1 77.4
- 5
DeepSeek V4.1 Flash ↑1 75.5
- 6 Kimi K3 ↑1 75.3
- 7 GPT-5.6 Sol ↓2 75.1
- 8
Gemini 3.8 Flash 68.9
- 9 GLM-5.3-Flash 68.5
- 10
Muse Spark 1.3 ↑1 67.8
Media
11 models we track have a score here.
- 1 GPT-6 Astra 75.6
- 2 GPT-6.1 Sol 74.3
- 3
Gemini 4 Argon 73.5
- 4 Claude Opus 5.5 71.6
- 5 GPT-5.6 Sol 67.1
- 6
DeepSeek V4.1 Flash 65.8
- 7 Kimi K3 62.5
- 8
Gemini 3.8 Flash 62.1
- 9 GLM-5.3-Flash 60.7
- 10
Doubao Seed 2.1 Pro 59.0
Video
11 models we track have a score here.
- 1 GPT-6 Astra 73.3
- 2 GPT-6.1 Sol 71.6
- 3
Gemini 4 Argon 70.7
- 4 Claude Opus 5.5 67.8
- 5 GPT-5.6 Sol 63.9
- 6
DeepSeek V4.1 Flash 62.3
- 7 Kimi K3 58.8
- 8
Gemini 3.8 Flash 57.2
- 9 GLM-5.3-Flash 56.5
- 10
Doubao Seed 2.1 Pro 54.7
Image
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 75.1
- 2 GPT-6 Astra ↓1 74.7
- 3 GPT-6.1 Sol ↓1 74.5
- 4
Gemini 4 Argon ↓1 73.7
- 5 GPT-5.6 Sol 68.3
- 6
DeepSeek V4.1 Flash 66.7
- 7
Gemini 3.8 Flash ↑1 65.4
- 8
Doubao Seed 2.1 Pro ↑2 63.6
- 9
Muse Spark 1.3 ↑2 62.6
- 10 GLM-5.3-Flash ↓1 62.3
Continuation
11 models we track have a score here.
- 1 GPT-6 Astra 85.4
- 2 GPT-6.1 Sol 83.8
- 3
Gemini 4 Argon 83.6
- 4 Claude Opus 5.5 80.4
- 5
DeepSeek V4.1 Flash ↑1 77.7
- 6 Kimi K3 ↑1 77.4
- 7 GPT-5.6 Sol ↓2 77.0
- 8
Gemini 3.8 Flash 74.5
- 9 GLM-5.3-Flash 73.6
- 10
Muse Spark 1.3 ↑1 69.9
Rule score (Text)
11 models we track have a score here.
- 1 GPT-6 Astra 85.6
- 2 GPT-6.1 Sol 84.4
- 3 Claude Opus 5.5 ↑1 83.0
- 4
Gemini 4 Argon ↓1 82.0
- 5 Kimi K3 ↑2 81.3
- 6 GPT-5.6 Sol ↓1 80.8
- 7
DeepSeek V4.1 Flash ↓1 80.1
- 8
Gemini 3.8 Flash 78.6
- 9
Muse Spark 1.3 ↑2 77.4
- 10 GLM-5.3-Flash ↓1 76.0
Gemini 4 judge (Text)
11 models we track have a score here.
- 1 GPT-6 Astra 76.5
- 2 GPT-6.1 Sol 73.3
- 3 Claude Opus 5.5 ↑1 73.0
- 4
Gemini 4 Argon ↓1 72.7
- 5
DeepSeek V4.1 Flash ↑1 70.8
- 6 GPT-5.6 Sol ↓1 69.4
- 7 Kimi K3 69.3
- 8 GLM-5.3-Flash ↑1 61.0
- 9
Gemini 3.8 Flash ↓1 59.1
- 10
Muse Spark 1.3 ↑1 58.3
Rule score (Media)
11 models we track have a score here.
- 1 GPT-6 Astra 60.1
- 2
Gemini 4 Argon ↑1 59.7
- 3 Claude Opus 5.5 ↑1 59.2
- 4 GPT-6.1 Sol ↓2 59.2
- 5 GPT-5.6 Sol 51.5
- 6
DeepSeek V4.1 Flash 49.3
- 7
Gemini 3.8 Flash ↑1 49.3
- 8 Kimi K3 ↓1 48.9
- 9 GLM-5.3-Flash 46.3
- 10
Doubao Seed 2.1 Pro 45.4
Gemini 4 judge (Media)
11 models we track have a score here.
- 1 GPT-6 Astra 91.0
- 2 GPT-6.1 Sol 89.3
- 3
Gemini 4 Argon 87.3
- 4 Claude Opus 5.5 84.1
- 5 GPT-5.6 Sol 82.7
- 6
DeepSeek V4.1 Flash 82.3
- 7 Kimi K3 76.1
- 8 GLM-5.3-Flash ↑1 75.0
- 9
Gemini 3.8 Flash ↓1 74.8
- 10
Doubao Seed 2.1 Pro 72.6
Rule score (Video)
11 models we track have a score here.
- 1 GPT-6 Astra 56.8
- 2 GPT-6.1 Sol 56.0
- 3
Gemini 4 Argon 55.8
- 4 Claude Opus 5.5 53.9
- 5 GPT-5.6 Sol 49.0
- 6
Gemini 3.8 Flash ↑2 46.2
- 7
DeepSeek V4.1 Flash ↓1 45.7
- 8 Kimi K3 ↓1 45.2
- 9 GLM-5.3-Flash 42.6
- 10
Doubao Seed 2.1 Pro 42.2
Gemini 4 judge (Video)
11 models we track have a score here.
- 1 GPT-6 Astra 89.8
- 2 GPT-6.1 Sol 87.3
- 3
Gemini 4 Argon 85.6
- 4 Claude Opus 5.5 81.6
- 5
DeepSeek V4.1 Flash ↑1 78.8
- 6 GPT-5.6 Sol ↓1 78.7
- 7 Kimi K3 72.5
- 8 GLM-5.3-Flash ↑1 70.2
- 9
Gemini 3.8 Flash ↓1 68.3
- 10
Doubao Seed 2.1 Pro 67.2
Rule score (Image)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 59.8
- 2
Gemini 4 Argon ↑1 57.1
- 3 GPT-6 Astra ↓2 55.8
- 4 GPT-6.1 Sol ↓2 55.3
- 5 GPT-5.6 Sol 44.2
- 6
DeepSeek V4.1 Flash 43.2
- 7
Gemini 3.8 Flash ↑1 43.1
- 8 Kimi K3 ↓1 42.4
- 9
Muse Spark 1.3 ↑2 40.5
- 10
Doubao Seed 2.1 Pro 40.4
Gemini 4 judge (Image)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 93.8
- 2 GPT-6 Astra ↓1 93.6
- 3 GPT-5.6 Sol ↑2 92.5
- 4 Claude Opus 5.5 90.4
- 5
Gemini 4 Argon ↓2 90.3
- 6
DeepSeek V4.1 Flash 90.2
- 7
Gemini 3.8 Flash ↑1 87.7
- 8
Doubao Seed 2.1 Pro ↑2 86.8
- 9 GLM-5.3-Flash 84.8
- 10
Muse Spark 1.3 ↑1 84.7
Rule score (Continuation)
11 models we track have a score here.
- 1 GPT-6 Astra 79.8
- 2
Gemini 4 Argon ↑1 78.5
- 3 Claude Opus 5.5 ↑1 78.2
- 4 GPT-6.1 Sol ↓2 78.1
- 5 Kimi K3 ↑2 73.9
- 6
DeepSeek V4.1 Flash 73.0
- 7 GPT-5.6 Sol ↓2 73.0
- 8
Gemini 3.8 Flash 71.4
- 9 GLM-5.3-Flash 71.0
- 10
Muse Spark 1.3 ↑1 67.2
Gemini 4 judge (Continuation)
11 models we track have a score here.
- 1 GPT-6 Astra 91.1
- 2 GPT-6.1 Sol 89.5
- 3
Gemini 4 Argon 88.8
- 4 Claude Opus 5.5 82.5
- 5
DeepSeek V4.1 Flash ↑1 82.4
- 6 GPT-5.6 Sol ↓1 81.1
- 7 Kimi K3 80.9
- 8
Gemini 3.8 Flash 77.6
- 9 GLM-5.3-Flash 76.3
- 10
Muse Spark 1.3 ↑1 72.5
Home
11 models we track have a score here.
- 1 GPT-6 Astra 78.0
- 2 GPT-6.1 Sol 77.7
- 3 Claude Opus 5.5 ↑1 75.6
- 4
Gemini 4 Argon ↓1 75.0
- 5 GPT-5.6 Sol 70.5
- 6
DeepSeek V4.1 Flash 68.5
- 7
Gemini 3.8 Flash ↑1 65.8
- 8 Kimi K3 ↓1 65.5
- 9
Doubao Seed 2.1 Pro ↑1 62.7
- 10
Muse Spark 1.3 ↑1 61.8
Public
11 models we track have a score here.
- 1 GPT-6 Astra 79.1
- 2 GPT-6.1 Sol 77.3
- 3 Claude Opus 5.5 ↑1 76.0
- 4
Gemini 4 Argon ↓1 75.8
- 5 GPT-5.6 Sol 71.9
- 6
DeepSeek V4.1 Flash 70.4
- 7 Kimi K3 67.3
- 8
Gemini 3.8 Flash 65.1
- 9 GLM-5.3-Flash 64.8
- 10
Muse Spark 1.3 ↑1 63.9
Retail
11 models we track have a score here.
- 1 GPT-6 Astra 81.2
- 2 Claude Opus 5.5 ↑2 77.1
- 3
Gemini 4 Argon 76.3
- 4 GPT-6.1 Sol ↓2 75.2
- 5
DeepSeek V4.1 Flash ↑1 73.0
- 6 GPT-5.6 Sol ↓1 71.6
- 7 Kimi K3 69.3
- 8 GLM-5.3-Flash ↑1 66.1
- 9
Gemini 3.8 Flash ↓1 65.4
- 10
Muse Spark 1.3 ↑1 65.0
Sport
11 models we track have a score here.
- 1 GPT-6 Astra 79.7
- 2 GPT-6.1 Sol 76.6
- 3
Gemini 4 Argon 75.2
- 4 Claude Opus 5.5 74.0
- 5
DeepSeek V4.1 Flash ↑1 72.4
- 6 Kimi K3 ↑1 71.1
- 7 GPT-5.6 Sol ↓2 71.1
- 8
Gemini 3.8 Flash 67.5
- 9 GLM-5.3-Flash 67.3
- 10
Muse Spark 1.3 ↑1 63.8
Transit
11 models we track have a score here.
- 1 GPT-6 Astra 76.5
- 2 GPT-6.1 Sol 76.1
- 3
Gemini 4 Argon 74.5
- 4 Claude Opus 5.5 73.5
- 5 GPT-5.6 Sol 70.2
- 6
DeepSeek V4.1 Flash 69.5
- 7 Kimi K3 68.5
- 8
Gemini 3.8 Flash 62.3
- 9 GLM-5.3-Flash 62.0
- 10
Doubao Seed 2.1 Pro 60.9
Street
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 73.6
- 2 GPT-6 Astra ↓1 73.3
- 3 GPT-6.1 Sol ↓1 72.5
- 4 Claude Opus 5.5 71.7
- 5 GPT-5.6 Sol 67.0
- 6
DeepSeek V4.1 Flash 65.2
- 7 Kimi K3 62.8
- 8
Gemini 3.8 Flash 60.4
- 9 GLM-5.3-Flash 59.6
- 10
Doubao Seed 2.1 Pro 57.7
Nature
11 models we track have a score here.
- 1 GPT-6 Astra 77.1
- 2 GPT-6.1 Sol 76.5
- 3
Gemini 4 Argon 75.8
- 4 Claude Opus 5.5 72.7
- 5 GPT-5.6 Sol 69.7
- 6
DeepSeek V4.1 Flash 68.5
- 7 Kimi K3 66.8
- 8 GLM-5.3-Flash ↑1 64.9
- 9
Gemini 3.8 Flash ↓1 64.7
- 10
Muse Spark 1.3 ↑1 60.9
Home living
11 models we track have a score here.
- 1 GPT-6 Astra 78.0
- 2 GPT-6.1 Sol 77.6
- 3 Claude Opus 5.5 ↑1 76.3
- 4
Gemini 4 Argon ↓1 75.3
- 5 GPT-5.6 Sol 70.2
- 6
DeepSeek V4.1 Flash 68.3
- 7 Kimi K3 66.1
- 8
Gemini 3.8 Flash 65.3
- 9
Doubao Seed 2.1 Pro ↑1 62.2
- 10
Muse Spark 1.3 ↑1 60.8
Home kitchen
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 82.2
- 2 GPT-6 Astra ↓1 81.2
- 3
Gemini 4 Argon 78.8
- 4 Claude Opus 5.5 78.7
- 5
DeepSeek V4.1 Flash ↑1 71.6
- 6 GPT-5.6 Sol ↓1 71.6
- 7 Kimi K3 71.2
- 8
Gemini 3.8 Flash 70.2
- 9 GLM-5.3-Flash 65.3
- 10
Doubao Seed 2.1 Pro 64.7
Home bedroom
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 78.6
- 2 GPT-6 Astra ↓1 77.5
- 3 Claude Opus 5.5 ↑1 75.5
- 4
Gemini 4 Argon ↓1 74.0
- 5 GPT-5.6 Sol 71.5
- 6
DeepSeek V4.1 Flash 68.4
- 7 Kimi K3 66.3
- 8
Gemini 3.8 Flash 66.0
- 9 GLM-5.3-Flash 65.2
- 10
Doubao Seed 2.1 Pro 64.8
Home hotel
11 models we track have a score here.
- 1 GPT-6 Astra 75.8
- 2 GPT-6.1 Sol 74.0
- 3
Gemini 4 Argon 72.1
- 4 Claude Opus 5.5 70.6
- 5 GPT-5.6 Sol 69.8
- 6
DeepSeek V4.1 Flash 67.4
- 7
Gemini 3.8 Flash ↑1 64.6
- 8
Muse Spark 1.3 ↑3 62.9
- 9 GLM-5.3-Flash 62.0
- 10
Doubao Seed 2.1 Pro 61.5
Public office
11 models we track have a score here.
- 1 GPT-6 Astra 78.3
- 2 GPT-6.1 Sol 77.7
- 3 Claude Opus 5.5 ↑1 76.3
- 4
Gemini 4 Argon ↓1 74.5
- 5 GPT-5.6 Sol 70.2
- 6
DeepSeek V4.1 Flash 68.2
- 7 GLM-5.3-Flash ↑2 66.8
- 8
Gemini 3.8 Flash 64.1
- 9
Doubao Seed 2.1 Pro ↑1 61.6
- 10 Kimi K3 ↓3 60.7
Public school
11 models we track have a score here.
- 1 GPT-6 Astra 77.1
- 2
Gemini 4 Argon ↑1 76.1
- 3 GPT-6.1 Sol ↓1 75.8
- 4 Claude Opus 5.5 73.8
- 5 GPT-5.6 Sol 71.3
- 6
DeepSeek V4.1 Flash 68.4
- 7 Kimi K3 67.4
- 8
Doubao Seed 2.1 Pro ↑2 65.8
- 9
Muse Spark 1.3 ↑2 64.6
- 10
Gemini 3.8 Flash ↓2 62.8
Public hospital
11 models we track have a score here.
- 1 GPT-6 Astra 79.7
- 2 Claude Opus 5.5 ↑2 78.0
- 3
DeepSeek V4.1 Flash ↑3 76.3
- 4 Kimi K3 ↑3 76.0
- 5 GPT-5.6 Sol 72.3
- 6
Gemini 4 Argon ↓3 69.4
- 7 GPT-6.1 Sol ↓5 68.5
- 8
Muse Spark 1.3 ↑3 65.6
- 9
Doubao Seed 2.1 Pro ↑1 64.4
- 10
Gemini 3.8 Flash ↓2 63.4
Public civic
11 models we track have a score here.
- 1 GPT-6 Astra 82.3
- 2 GPT-6.1 Sol 81.8
- 3
Gemini 4 Argon 80.2
- 4 Claude Opus 5.5 76.8
- 5 GPT-5.6 Sol 75.3
- 6 Kimi K3 ↑1 75.0
- 7
DeepSeek V4.1 Flash ↓1 73.7
- 8
Gemini 3.8 Flash 70.1
- 9 GLM-5.3-Flash 69.9
- 10
Muse Spark 1.3 ↑1 68.1
Retail store
11 models we track have a score here.
- 1 GPT-6 Astra 78.8
- 2
Gemini 4 Argon ↑1 74.7
- 3
DeepSeek V4.1 Flash ↑3 73.1
- 4 Claude Opus 5.5 72.9
- 5 GPT-6.1 Sol ↓3 71.7
- 6 GPT-5.6 Sol ↓1 71.1
- 7 Kimi K3 69.3
- 8
Doubao Seed 2.1 Pro ↑2 64.7
- 9
Gemini 3.8 Flash ↓1 62.1
- 10 GLM-5.3-Flash ↓1 60.1
Retail food
11 models we track have a score here.
- 1 GPT-6 Astra 83.8
- 2 GPT-6.1 Sol 80.5
- 3 Claude Opus 5.5 ↑1 79.0
- 4
Gemini 4 Argon ↓1 76.7
- 5
DeepSeek V4.1 Flash ↑1 74.0
- 6 GLM-5.3-Flash ↑3 71.5
- 7
Muse Spark 1.3 ↑4 71.3
- 8 GPT-5.6 Sol ↓3 68.5
- 9
Gemini 3.8 Flash ↓1 66.4
- 10 Kimi K3 ↓3 66.3
Retail leisure
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 82.3
- 2 GPT-6 Astra ↓1 81.7
- 3
Gemini 4 Argon 78.9
- 4 GPT-5.6 Sol ↑1 77.9
- 5 Kimi K3 ↑2 74.3
- 6 GPT-6.1 Sol ↓4 73.5
- 7
DeepSeek V4.1 Flash ↓1 71.4
- 8
Gemini 3.8 Flash 70.4
- 9 GLM-5.3-Flash 69.1
- 10
Muse Spark 1.3 ↑1 65.9
Sport ball
11 models we track have a score here.
- 1 GPT-6 Astra 83.0
- 2 GPT-6.1 Sol 82.4
- 3
Gemini 4 Argon 81.3
- 4 Claude Opus 5.5 77.6
- 5
DeepSeek V4.1 Flash ↑1 75.0
- 6 Kimi K3 ↑1 74.6
- 7 GPT-5.6 Sol ↓2 73.8
- 8
Gemini 3.8 Flash 72.1
- 9 GLM-5.3-Flash 70.0
- 10
Muse Spark 1.3 ↑1 67.3
Sport board
11 models we track have a score here.
- 1 GPT-6 Astra 76.9
- 2 GPT-6.1 Sol 71.2
- 3
DeepSeek V4.1 Flash ↑3 70.7
- 4 Kimi K3 ↑3 69.3
- 5
Gemini 4 Argon ↓2 68.6
- 6 Claude Opus 5.5 ↓2 68.3
- 7 GLM-5.3-Flash ↑2 67.1
- 8 GPT-5.6 Sol ↓3 66.0
- 9
Doubao Seed 2.1 Pro ↑1 64.7
- 10
Muse Spark 1.3 ↑1 64.5
Sport snow
11 models we track have a score here.
- 1 GPT-6 Astra 82.0
- 2
Gemini 4 Argon ↑1 79.9
- 3 GPT-6.1 Sol ↓1 76.7
- 4 Kimi K3 ↑3 76.2
- 5
DeepSeek V4.1 Flash ↑1 76.0
- 6 Claude Opus 5.5 ↓2 74.0
- 7 GPT-5.6 Sol ↓2 72.8
- 8
Doubao Seed 2.1 Pro ↑2 70.1
- 9
Muse Spark 1.3 ↑2 69.7
- 10 GLM-5.3-Flash ↓1 69.4
Sport motor
11 models we track have a score here.
- 1 GPT-6 Astra 78.3
- 2 Claude Opus 5.5 ↑2 76.5
- 3 GPT-6.1 Sol ↓1 75.5
- 4 GPT-5.6 Sol ↑1 72.5
- 5
DeepSeek V4.1 Flash ↑1 72.2
- 6 Kimi K3 ↑1 71.0
- 7
Gemini 3.8 Flash ↑1 69.4
- 8
Gemini 4 Argon ↓5 65.4
- 9 GLM-5.3-Flash 60.6
- 10
Doubao Seed 2.1 Pro 57.5
Sport water
11 models we track have a score here.
- 1 GPT-6 Astra 80.1
- 2
Gemini 4 Argon ↑1 78.5
- 3 Kimi K3 ↑4 78.4
- 4 GPT-6.1 Sol ↓2 77.0
- 5
DeepSeek V4.1 Flash ↑1 75.5
- 6 GPT-5.6 Sol ↓1 75.1
- 7 Claude Opus 5.5 ↓3 71.2
- 8 GLM-5.3-Flash ↑1 70.8
- 9
Gemini 3.8 Flash ↓1 70.6
- 10
Muse Spark 1.3 ↑1 69.2
Sport athletics
11 models we track have a score here.
- 1 GPT-6 Astra 81.7
- 2
Gemini 4 Argon ↑1 81.1
- 3 Claude Opus 5.5 ↑1 78.8
- 4 GPT-6.1 Sol ↓2 78.3
- 5 GPT-5.6 Sol 75.6
- 6
DeepSeek V4.1 Flash 74.0
- 7
Gemini 3.8 Flash ↑1 69.0
- 8 GLM-5.3-Flash ↑1 67.8
- 9 Kimi K3 ↓2 67.6
- 10
Muse Spark 1.3 ↑1 64.4
Sport animal
11 models we track have a score here.
- 1 GPT-6 Astra 74.8
- 2 GPT-6.1 Sol 71.6
- 3
Gemini 4 Argon 68.1
- 4 Claude Opus 5.5 67.0
- 5
DeepSeek V4.1 Flash ↑1 63.9
- 6 GLM-5.3-Flash ↑3 63.3
- 7 GPT-5.6 Sol ↓2 63.1
- 8 Kimi K3 ↓1 62.8
- 9
Gemini 3.8 Flash ↓1 61.0
- 10
Doubao Seed 2.1 Pro 60.5
Sport obstacle
11 models we track have a score here.
- 1 GPT-6 Astra 78.7
- 2
Gemini 4 Argon ↑1 76.3
- 3 GPT-6.1 Sol ↓1 76.3
- 4 Claude Opus 5.5 76.2
- 5
DeepSeek V4.1 Flash ↑1 70.0
- 6 GPT-5.6 Sol ↓1 69.8
- 7 Kimi K3 69.1
- 8 GLM-5.3-Flash ↑1 66.8
- 9
Gemini 3.8 Flash ↓1 66.0
- 10
Doubao Seed 2.1 Pro 61.4
Transit station
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 78.5
- 2 GPT-6 Astra ↓1 78.1
- 3 GPT-6.1 Sol ↓1 77.2
- 4 Claude Opus 5.5 74.3
- 5 GPT-5.6 Sol 73.0
- 6
DeepSeek V4.1 Flash 72.6
- 7 Kimi K3 71.9
- 8 GLM-5.3-Flash ↑1 63.8
- 9
Gemini 3.8 Flash ↓1 62.9
- 10
Doubao Seed 2.1 Pro 59.7
Transit airport
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 83.1
- 2 GPT-6.1 Sol 81.9
- 3 Claude Opus 5.5 ↑1 80.1
- 4 Kimi K3 ↑3 76.5
- 5 GPT-5.6 Sol 76.4
- 6 GLM-5.3-Flash ↑3 75.5
- 7 GPT-6 Astra ↓6 74.7
- 8
DeepSeek V4.1 Flash ↓2 73.8
- 9
Gemini 3.8 Flash ↓1 72.7
- 10
Doubao Seed 2.1 Pro 65.0
Transit port
11 models we track have a score here.
- 1 GPT-6 Astra 78.3
- 2 GPT-6.1 Sol 77.2
- 3
Gemini 4 Argon 75.0
- 4 Claude Opus 5.5 74.3
- 5 GPT-5.6 Sol 71.8
- 6 Kimi K3 ↑1 69.8
- 7
DeepSeek V4.1 Flash ↓1 68.7
- 8 GLM-5.3-Flash ↑1 67.9
- 9
Gemini 3.8 Flash ↓1 62.8
- 10
Doubao Seed 2.1 Pro 61.7
Transit industrial
11 models we track have a score here.
- 1 GPT-6 Astra 75.3
- 2 GPT-6.1 Sol 74.5
- 3 Claude Opus 5.5 ↑1 72.0
- 4
Gemini 4 Argon ↓1 70.9
- 5
DeepSeek V4.1 Flash ↑1 67.4
- 6 GPT-5.6 Sol ↓1 67.3
- 7 Kimi K3 65.2
- 8
Doubao Seed 2.1 Pro ↑2 60.8
- 9
Gemini 3.8 Flash ↓1 60.6
- 10 GLM-5.3-Flash ↓1 57.4
Street city
11 models we track have a score here.
- 1 GPT-6 Astra 72.8
- 2 Claude Opus 5.5 ↑2 71.9
- 3
Gemini 4 Argon 71.0
- 4 GPT-6.1 Sol ↓2 69.2
- 5
DeepSeek V4.1 Flash ↑1 65.7
- 6 GPT-5.6 Sol ↓1 64.3
- 7 Kimi K3 61.1
- 8 GLM-5.3-Flash ↑1 59.8
- 9
Gemini 3.8 Flash ↓1 58.4
- 10
Doubao Seed 2.1 Pro 55.5
Street residential
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 76.9
- 2 Claude Opus 5.5 ↑2 75.5
- 3 GPT-6 Astra ↓2 74.0
- 4 GPT-6.1 Sol ↓2 73.7
- 5 GPT-5.6 Sol 71.2
- 6 Kimi K3 ↑1 67.9
- 7
DeepSeek V4.1 Flash ↓1 65.3
- 8
Gemini 3.8 Flash 65.1
- 9 GLM-5.3-Flash 64.0
- 10
Doubao Seed 2.1 Pro 62.2
Street old town
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 75.0
- 2 GPT-6 Astra ↓1 74.0
- 3
Gemini 4 Argon 73.0
- 4 Claude Opus 5.5 67.7
- 5 GPT-5.6 Sol 65.8
- 6
DeepSeek V4.1 Flash 63.8
- 7 Kimi K3 60.6
- 8
Doubao Seed 2.1 Pro ↑2 57.3
- 9
Gemini 3.8 Flash ↓1 55.9
- 10 GLM-5.3-Flash ↓1 55.4
Street rooftop
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 75.7
- 2 GPT-6.1 Sol 74.6
- 3 GPT-6 Astra ↓2 70.4
- 4 Claude Opus 5.5 70.0
- 5 GPT-5.6 Sol 68.1
- 6
Gemini 3.8 Flash ↑2 67.1
- 7
DeepSeek V4.1 Flash ↓1 66.8
- 8 Kimi K3 ↓1 59.2
- 9 GLM-5.3-Flash 57.3
- 10
Muse Spark 1.3 ↑1 52.8
Nature snow
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 76.2
- 2 GPT-6 Astra ↓1 75.3
- 3
Gemini 4 Argon 73.2
- 4 Claude Opus 5.5 69.3
- 5 GPT-5.6 Sol 66.8
- 6
DeepSeek V4.1 Flash 64.6
- 7
Gemini 3.8 Flash ↑1 63.5
- 8 Kimi K3 ↓1 62.6
- 9 GLM-5.3-Flash 62.4
- 10
Muse Spark 1.3 ↑1 59.9
Nature mountain
11 models we track have a score here.
- 1 GPT-6 Astra 76.1
- 2 GPT-6.1 Sol 75.6
- 3
Gemini 4 Argon 75.2
- 4 Claude Opus 5.5 69.8
- 5 GPT-5.6 Sol 69.2
- 6
DeepSeek V4.1 Flash 67.2
- 7
Gemini 3.8 Flash ↑1 63.4
- 8 Kimi K3 ↓1 63.0
- 9 GLM-5.3-Flash 62.9
- 10
Doubao Seed 2.1 Pro 58.5
Nature water
11 models we track have a score here.
- 1 GPT-6 Astra 79.8
- 2 Claude Opus 5.5 ↑2 79.2
- 3
Gemini 4 Argon 78.7
- 4 GPT-6.1 Sol ↓2 77.8
- 5 Kimi K3 ↑2 75.3
- 6
DeepSeek V4.1 Flash 73.4
- 7 GPT-5.6 Sol ↓2 72.6
- 8 GLM-5.3-Flash ↑1 69.4
- 9
Gemini 3.8 Flash ↓1 67.3
- 10
Muse Spark 1.3 ↑1 66.5
Rule score (Home)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 65.6
- 2 GPT-6.1 Sol 64.9
- 3 GPT-6 Astra ↓2 64.7
- 4
Gemini 4 Argon ↓1 64.7
- 5 GPT-5.6 Sol 55.9
- 6
Gemini 3.8 Flash ↑2 54.6
- 7
DeepSeek V4.1 Flash ↓1 54.4
- 8 Kimi K3 ↓1 52.8
- 9
Doubao Seed 2.1 Pro ↑1 50.5
- 10
Muse Spark 1.3 ↑1 48.7
Gemini 4 judge (Home)
11 models we track have a score here.
- 1 GPT-6 Astra 91.2
- 2 GPT-6.1 Sol 90.5
- 3 Claude Opus 5.5 ↑1 85.5
- 4
Gemini 4 Argon ↓1 85.2
- 5 GPT-5.6 Sol 85.0
- 6
DeepSeek V4.1 Flash 82.6
- 7 Kimi K3 78.1
- 8
Gemini 3.8 Flash 77.0
- 9
Doubao Seed 2.1 Pro ↑1 75.0
- 10
Muse Spark 1.3 ↑1 74.8
Rule score (Public)
11 models we track have a score here.
- 1 GPT-6 Astra 73.3
- 2 Claude Opus 5.5 ↑2 72.2
- 3
Gemini 4 Argon 72.1
- 4 GPT-6.1 Sol ↓2 71.1
- 5 GPT-5.6 Sol 64.7
- 6
DeepSeek V4.1 Flash 63.1
- 7
Gemini 3.8 Flash ↑1 61.1
- 8 Kimi K3 ↓1 60.9
- 9
Muse Spark 1.3 ↑2 60.2
- 10 GLM-5.3-Flash ↓1 59.6
Gemini 4 judge (Public)
11 models we track have a score here.
- 1 GPT-6 Astra 85.0
- 2 GPT-6.1 Sol 83.5
- 3 Claude Opus 5.5 ↑1 79.8
- 4
Gemini 4 Argon ↓1 79.4
- 5 GPT-5.6 Sol 79.1
- 6
DeepSeek V4.1 Flash 77.7
- 7 Kimi K3 73.8
- 8 GLM-5.3-Flash ↑1 70.0
- 9
Gemini 3.8 Flash ↓1 69.2
- 10
Doubao Seed 2.1 Pro 68.6
Rule score (Retail)
11 models we track have a score here.
- 1 GPT-6 Astra 79.0
- 2
Gemini 4 Argon ↑1 78.4
- 3 Claude Opus 5.5 ↑1 77.2
- 4 GPT-6.1 Sol ↓2 76.0
- 5 GPT-5.6 Sol 72.0
- 6
DeepSeek V4.1 Flash 71.7
- 7 Kimi K3 71.5
- 8
Gemini 3.8 Flash 70.1
- 9
Muse Spark 1.3 ↑2 68.9
- 10 GLM-5.3-Flash ↓1 67.4
Gemini 4 judge (Retail)
11 models we track have a score here.
- 1 GPT-6 Astra 83.4
- 2 Claude Opus 5.5 ↑2 77.0
- 3 GPT-6.1 Sol ↓1 74.4
- 4
DeepSeek V4.1 Flash ↑2 74.3
- 5
Gemini 4 Argon ↓2 74.2
- 6 GPT-5.6 Sol ↓1 71.2
- 7 Kimi K3 67.1
- 8 GLM-5.3-Flash ↑1 64.8
- 9
Muse Spark 1.3 ↑2 61.1
- 10
Gemini 3.8 Flash ↓2 60.7
Rule score (Sport)
11 models we track have a score here.
- 1 GPT-6 Astra 75.4
- 2 GPT-6.1 Sol 72.8
- 3 Claude Opus 5.5 ↑1 72.6
- 4
Gemini 4 Argon ↓1 70.3
- 5 GPT-5.6 Sol 68.0
- 6 Kimi K3 ↑1 67.6
- 7
DeepSeek V4.1 Flash ↓1 67.2
- 8
Gemini 3.8 Flash 65.8
- 9 GLM-5.3-Flash 65.0
- 10
Muse Spark 1.3 ↑1 63.7
Gemini 4 judge (Sport)
11 models we track have a score here.
- 1 GPT-6 Astra 84.0
- 2 GPT-6.1 Sol 80.4
- 3
Gemini 4 Argon 80.1
- 4
DeepSeek V4.1 Flash ↑2 77.5
- 5 Claude Opus 5.5 ↓1 75.5
- 6 Kimi K3 ↑1 74.6
- 7 GPT-5.6 Sol ↓2 74.1
- 8 GLM-5.3-Flash ↑1 69.7
- 9
Gemini 3.8 Flash ↓1 69.3
- 10
Doubao Seed 2.1 Pro 65.3
Rule score (Transit)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 73.9
- 2 GPT-6 Astra ↓1 73.2
- 3
Gemini 4 Argon 71.1
- 4 Claude Opus 5.5 69.9
- 5 GPT-5.6 Sol 67.5
- 6 Kimi K3 ↑1 67.2
- 7
DeepSeek V4.1 Flash ↓1 66.5
- 8
Gemini 3.8 Flash 64.6
- 9 GLM-5.3-Flash 62.1
- 10
Doubao Seed 2.1 Pro 61.4
Gemini 4 judge (Transit)
11 models we track have a score here.
- 1 GPT-6 Astra 79.9
- 2 GPT-6.1 Sol 78.3
- 3
Gemini 4 Argon 77.8
- 4 Claude Opus 5.5 77.1
- 5 GPT-5.6 Sol 72.8
- 6
DeepSeek V4.1 Flash 72.4
- 7 Kimi K3 69.9
- 8 GLM-5.3-Flash ↑1 61.8
- 9
Doubao Seed 2.1 Pro ↑1 60.4
- 10
Gemini 3.8 Flash ↓2 60.1
Rule score (Street)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 59.8
- 2 Claude Opus 5.5 ↑2 57.4
- 3 GPT-6 Astra ↓2 56.9
- 4 GPT-6.1 Sol ↓2 56.9
- 5 GPT-5.6 Sol 51.5
- 6 Kimi K3 ↑1 50.1
- 7
Gemini 3.8 Flash ↑1 49.8
- 8
DeepSeek V4.1 Flash ↓2 47.6
- 9 GLM-5.3-Flash 46.4
- 10
Doubao Seed 2.1 Pro 45.2
Gemini 4 judge (Street)
11 models we track have a score here.
- 1 GPT-6 Astra 89.7
- 2 GPT-6.1 Sol 88.0
- 3
Gemini 4 Argon 87.4
- 4 Claude Opus 5.5 85.9
- 5
DeepSeek V4.1 Flash ↑1 82.7
- 6 GPT-5.6 Sol ↓1 82.5
- 7 Kimi K3 75.5
- 8 GLM-5.3-Flash ↑1 72.9
- 9
Gemini 3.8 Flash ↓1 71.0
- 10
Doubao Seed 2.1 Pro 70.2
Rule score (Nature)
11 models we track have a score here.
- 1 GPT-6 Astra 72.2
- 2 GPT-6.1 Sol 71.8
- 3
Gemini 4 Argon 69.0
- 4 Claude Opus 5.5 68.5
- 5 Kimi K3 ↑2 65.3
- 6 GPT-5.6 Sol ↓1 65.2
- 7
DeepSeek V4.1 Flash ↓1 63.5
- 8
Gemini 3.8 Flash 63.4
- 9 GLM-5.3-Flash 61.8
- 10
Muse Spark 1.3 ↑1 60.9
Gemini 4 judge (Nature)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 82.7
- 2 GPT-6 Astra ↓1 82.0
- 3 GPT-6.1 Sol ↓1 81.2
- 4 Claude Opus 5.5 77.0
- 5 GPT-5.6 Sol 74.2
- 6
DeepSeek V4.1 Flash 73.6
- 7 Kimi K3 68.4
- 8 GLM-5.3-Flash ↑1 68.0
- 9
Gemini 3.8 Flash ↓1 65.9
- 10
Muse Spark 1.3 ↑1 61.0
Rule score (Home living)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 66.8
- 2
Gemini 4 Argon ↑1 66.8
- 3 GPT-6.1 Sol ↓1 66.7
- 4 GPT-6 Astra ↓3 66.6
- 5 GPT-5.6 Sol 58.5
- 6
DeepSeek V4.1 Flash 56.4
- 7
Gemini 3.8 Flash ↑1 56.2
- 8 Kimi K3 ↓1 54.0
- 9
Doubao Seed 2.1 Pro ↑1 52.3
- 10
Muse Spark 1.3 ↑1 50.0
Gemini 4 judge (Home living)
11 models we track have a score here.
- 1 GPT-6 Astra 89.5
- 2 GPT-6.1 Sol 88.6
- 3 Claude Opus 5.5 ↑1 85.9
- 4
Gemini 4 Argon ↓1 83.8
- 5 GPT-5.6 Sol 81.9
- 6
DeepSeek V4.1 Flash 80.2
- 7 Kimi K3 78.2
- 8
Gemini 3.8 Flash 74.5
- 9
Doubao Seed 2.1 Pro ↑1 72.0
- 10
Muse Spark 1.3 ↑1 71.5
Rule score (Home kitchen)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 66.7
- 2 GPT-6 Astra ↓1 65.7
- 3
Gemini 4 Argon 65.0
- 4 Claude Opus 5.5 64.5
- 5 Kimi K3 ↑2 55.3
- 6
Gemini 3.8 Flash ↑2 52.7
- 7 GLM-5.3-Flash ↑2 52.2
- 8 GPT-5.6 Sol ↓3 52.1
- 9
DeepSeek V4.1 Flash ↓3 51.4
- 10
Doubao Seed 2.1 Pro 50.9
Gemini 4 judge (Home kitchen)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 97.7
- 2 GPT-6 Astra ↓1 96.7
- 3 Claude Opus 5.5 ↑1 92.9
- 4
Gemini 4 Argon ↓1 92.6
- 5
DeepSeek V4.1 Flash ↑1 91.8
- 6 GPT-5.6 Sol ↓1 91.1
- 7
Gemini 3.8 Flash ↑1 87.8
- 8 Kimi K3 ↓1 87.1
- 9
Muse Spark 1.3 ↑2 81.7
- 10
Doubao Seed 2.1 Pro 78.5
Rule score (Home bedroom)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 67.6
- 2 GPT-6.1 Sol 62.4
- 3
Gemini 4 Argon 61.6
- 4 GPT-6 Astra ↓3 60.9
- 5 GPT-5.6 Sol 51.8
- 6
Gemini 3.8 Flash ↑2 50.7
- 7 Kimi K3 50.2
- 8
DeepSeek V4.1 Flash ↓2 50.0
- 9 GLM-5.3-Flash 47.6
- 10
Muse Spark 1.3 ↑1 47.3
Gemini 4 judge (Home bedroom)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 94.8
- 2 GPT-6 Astra ↓1 94.2
- 3 GPT-5.6 Sol ↑2 91.2
- 4
DeepSeek V4.1 Flash ↑2 86.7
- 5
Gemini 4 Argon ↓2 86.4
- 6
Doubao Seed 2.1 Pro ↑4 84.0
- 7 Claude Opus 5.5 ↓3 83.3
- 8 GLM-5.3-Flash ↑1 82.8
- 9 Kimi K3 ↓2 82.5
- 10
Gemini 3.8 Flash ↓2 81.3
Rule score (Home hotel)
11 models we track have a score here.
- 1 GPT-6 Astra 60.2
- 2 Claude Opus 5.5 ↑2 59.5
- 3
Gemini 4 Argon 59.4
- 4 GPT-6.1 Sol ↓2 58.9
- 5
Gemini 3.8 Flash ↑3 53.4
- 6
DeepSeek V4.1 Flash 52.8
- 7 GPT-5.6 Sol ↓2 52.2
- 8 Kimi K3 ↓1 49.1
- 9
Doubao Seed 2.1 Pro ↑1 47.8
- 10
Muse Spark 1.3 ↑1 47.6
Gemini 4 judge (Home hotel)
11 models we track have a score here.
- 1 GPT-6 Astra 91.4
- 2 GPT-6.1 Sol 89.2
- 3 GPT-5.6 Sol ↑2 87.4
- 4
Gemini 4 Argon ↓1 84.7
- 5
DeepSeek V4.1 Flash ↑1 82.0
- 6 Claude Opus 5.5 ↓2 81.6
- 7 GLM-5.3-Flash ↑2 79.2
- 8
Muse Spark 1.3 ↑3 78.3
- 9
Gemini 3.8 Flash ↓1 75.9
- 10
Doubao Seed 2.1 Pro 75.2
Rule score (Public office)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 68.3
- 2 GPT-6 Astra ↓1 68.2
- 3 GPT-6.1 Sol ↓1 68.2
- 4
Gemini 4 Argon ↓1 66.4
- 5 GPT-5.6 Sol 60.1
- 6
DeepSeek V4.1 Flash 56.9
- 7
Gemini 3.8 Flash ↑1 56.7
- 8 GLM-5.3-Flash ↑1 55.5
- 9
Muse Spark 1.3 ↑2 55.0
- 10
Doubao Seed 2.1 Pro 52.2
Gemini 4 judge (Public office)
11 models we track have a score here.
- 1 GPT-6 Astra 88.4
- 2 GPT-6.1 Sol 87.3
- 3 Claude Opus 5.5 ↑1 84.4
- 4
Gemini 4 Argon ↓1 82.7
- 5 GPT-5.6 Sol 80.3
- 6
DeepSeek V4.1 Flash 79.6
- 7 GLM-5.3-Flash ↑2 78.1
- 8
Gemini 3.8 Flash 71.5
- 9
Doubao Seed 2.1 Pro ↑1 71.0
- 10 Kimi K3 ↓3 70.9
Rule score (Public school)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 65.5
- 2 GPT-6 Astra ↓1 65.3
- 3 GPT-6.1 Sol ↓1 63.9
- 4 Claude Opus 5.5 62.8
- 5 GPT-5.6 Sol 55.0
- 6 Kimi K3 ↑1 53.1
- 7
Doubao Seed 2.1 Pro ↑3 52.3
- 8
DeepSeek V4.1 Flash ↓2 51.0
- 9
Muse Spark 1.3 ↑2 50.3
- 10
Gemini 3.8 Flash ↓2 46.0
Gemini 4 judge (Public school)
11 models we track have a score here.
- 1 GPT-6 Astra 88.9
- 2 GPT-5.6 Sol ↑3 87.7
- 3 GPT-6.1 Sol ↓1 87.7
- 4
Gemini 4 Argon ↓1 86.7
- 5
DeepSeek V4.1 Flash ↑1 85.8
- 6 Claude Opus 5.5 ↓2 84.8
- 7 Kimi K3 81.6
- 8
Gemini 3.8 Flash 79.5
- 9
Doubao Seed 2.1 Pro ↑1 79.4
- 10
Muse Spark 1.3 ↑1 78.9
Rule score (Public hospital)
11 models we track have a score here.
- 1 GPT-6 Astra 91.1
- 2
Gemini 4 Argon ↑1 87.7
- 3 Claude Opus 5.5 ↑1 87.2
- 4 Kimi K3 ↑3 85.3
- 5
Muse Spark 1.3 ↑6 85.0
- 6
DeepSeek V4.1 Flash 84.3
- 7
Gemini 3.8 Flash ↑1 82.9
- 8 GPT-5.6 Sol ↓3 79.4
- 9
Doubao Seed 2.1 Pro ↑1 79.1
- 10 GPT-6.1 Sol ↓8 77.2
Gemini 4 judge (Public hospital)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 68.7
- 2 GPT-6 Astra ↓1 68.3
- 3
DeepSeek V4.1 Flash ↑3 68.3
- 4 Kimi K3 ↑3 66.6
- 5 GPT-5.6 Sol 65.2
- 6 GPT-6.1 Sol ↓4 59.7
- 7
Gemini 4 Argon ↓4 51.2
- 8
Doubao Seed 2.1 Pro ↑2 49.7
- 9 GLM-5.3-Flash 49.1
- 10
Muse Spark 1.3 ↑1 46.3
Rule score (Public civic)
11 models we track have a score here.
- 1 GPT-6 Astra 82.4
- 2
Gemini 4 Argon ↑1 82.0
- 3 Claude Opus 5.5 ↑1 81.9
- 4 GPT-6.1 Sol ↓2 80.8
- 5
DeepSeek V4.1 Flash ↑1 76.9
- 6 GPT-5.6 Sol ↓1 76.1
- 7 Kimi K3 76.0
- 8
Gemini 3.8 Flash 74.3
- 9 GLM-5.3-Flash 73.1
- 10
Muse Spark 1.3 ↑1 68.6
Gemini 4 judge (Public civic)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 82.9
- 2 GPT-6 Astra ↓1 82.2
- 3
Gemini 4 Argon 78.4
- 4 GPT-5.6 Sol ↑1 74.5
- 5 Kimi K3 ↑2 74.0
- 6 Claude Opus 5.5 ↓2 71.6
- 7
DeepSeek V4.1 Flash ↓1 70.5
- 8
Muse Spark 1.3 ↑3 67.6
- 9 GLM-5.3-Flash 66.8
- 10
Gemini 3.8 Flash ↓2 65.8
Rule score (Retail store)
11 models we track have a score here.
- 1 GPT-6 Astra 74.9
- 2
Gemini 4 Argon ↑1 73.1
- 3 Claude Opus 5.5 ↑1 71.8
- 4 GPT-6.1 Sol ↓2 71.0
- 5 GPT-5.6 Sol 68.3
- 6
DeepSeek V4.1 Flash 67.4
- 7 Kimi K3 66.8
- 8
Gemini 3.8 Flash 65.5
- 9
Doubao Seed 2.1 Pro ↑1 65.0
- 10
Muse Spark 1.3 ↑1 60.9
Gemini 4 judge (Retail store)
11 models we track have a score here.
- 1 GPT-6 Astra 82.7
- 2
DeepSeek V4.1 Flash ↑4 78.7
- 3
Gemini 4 Argon 76.2
- 4 Claude Opus 5.5 74.0
- 5 GPT-5.6 Sol 73.9
- 6 GPT-6.1 Sol ↓4 72.3
- 7 Kimi K3 71.7
- 8
Doubao Seed 2.1 Pro ↑2 64.3
- 9 GLM-5.3-Flash 62.2
- 10
Gemini 3.8 Flash ↓2 58.7
Rule score (Retail food)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 79.7
- 2 Claude Opus 5.5 ↑2 78.2
- 3 GPT-6 Astra ↓2 77.7
- 4 GPT-6.1 Sol ↓2 74.0
- 5
Muse Spark 1.3 ↑6 74.0
- 6 GLM-5.3-Flash ↑3 71.5
- 7 Kimi K3 70.6
- 8
Gemini 3.8 Flash 70.6
- 9
DeepSeek V4.1 Flash ↓3 70.0
- 10 GPT-5.6 Sol ↓5 69.6
Gemini 4 judge (Retail food)
11 models we track have a score here.
- 1 GPT-6 Astra 89.8
- 2 GPT-6.1 Sol 87.1
- 3 Claude Opus 5.5 ↑1 79.8
- 4
DeepSeek V4.1 Flash ↑2 77.9
- 5
Gemini 4 Argon ↓2 73.8
- 6 GLM-5.3-Flash ↑3 71.5
- 7
Muse Spark 1.3 ↑4 68.7
- 8 GPT-5.6 Sol ↓3 67.3
- 9
Gemini 3.8 Flash ↓1 62.2
- 10 Kimi K3 ↓3 61.9
Rule score (Retail leisure)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 89.5
- 2 GPT-6 Astra ↓1 89.2
- 3
Gemini 4 Argon 86.7
- 4 Claude Opus 5.5 86.4
- 5 GPT-5.6 Sol 83.3
- 6
DeepSeek V4.1 Flash 83.1
- 7 Kimi K3 82.3
- 8 GLM-5.3-Flash ↑1 79.5
- 9
Gemini 3.8 Flash ↓1 78.3
- 10
Muse Spark 1.3 ↑1 76.4
Gemini 4 judge (Retail leisure)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 78.2
- 2 GPT-6 Astra ↓1 74.1
- 3 GPT-5.6 Sol ↑2 72.5
- 4
Gemini 4 Argon ↓1 71.1
- 5 Kimi K3 ↑2 66.4
- 6
Gemini 3.8 Flash ↑2 62.4
- 7
DeepSeek V4.1 Flash ↓1 59.7
- 8 GLM-5.3-Flash ↑1 58.8
- 9 GPT-6.1 Sol ↓7 57.5
- 10
Muse Spark 1.3 ↑1 55.4
Rule score (Sport ball)
11 models we track have a score here.
- 1 GPT-6 Astra 80.2
- 2 GPT-6.1 Sol 79.3
- 3 Claude Opus 5.5 ↑1 78.7
- 4
Gemini 4 Argon ↓1 77.5
- 5 Kimi K3 ↑2 73.9
- 6 GPT-5.6 Sol ↓1 73.1
- 7
Gemini 3.8 Flash ↑1 72.2
- 8
DeepSeek V4.1 Flash ↓2 71.1
- 9
Muse Spark 1.3 ↑2 69.5
- 10 GLM-5.3-Flash ↓1 67.4
Gemini 4 judge (Sport ball)
11 models we track have a score here.
- 1 GPT-6 Astra 85.7
- 2 GPT-6.1 Sol 85.4
- 3
Gemini 4 Argon 85.1
- 4
DeepSeek V4.1 Flash ↑2 79.0
- 5 Claude Opus 5.5 ↓1 76.5
- 6 Kimi K3 ↑1 75.3
- 7 GPT-5.6 Sol ↓2 74.5
- 8 GLM-5.3-Flash ↑1 72.5
- 9
Gemini 3.8 Flash ↓1 72.0
- 10
Muse Spark 1.3 ↑1 65.1
Rule score (Sport board)
11 models we track have a score here.
- 1 GPT-6 Astra 70.1
- 2 GPT-6.1 Sol 68.5
- 3 Claude Opus 5.5 ↑1 67.4
- 4 GPT-5.6 Sol ↑1 63.7
- 5
DeepSeek V4.1 Flash ↑1 61.3
- 6 Kimi K3 ↑1 61.3
- 7
Gemini 4 Argon ↓4 59.8
- 8
Muse Spark 1.3 ↑3 59.5
- 9
Gemini 3.8 Flash ↓1 59.0
- 10 GLM-5.3-Flash ↓1 57.9
Gemini 4 judge (Sport board)
11 models we track have a score here.
- 1 GPT-6 Astra 83.7
- 2
DeepSeek V4.1 Flash ↑4 80.0
- 3
Gemini 4 Argon 77.4
- 4 Kimi K3 ↑3 77.3
- 5 GLM-5.3-Flash ↑4 76.4
- 6 GPT-6.1 Sol ↓4 74.0
- 7
Doubao Seed 2.1 Pro ↑3 73.8
- 8
Muse Spark 1.3 ↑3 69.5
- 9 Claude Opus 5.5 ↓5 69.1
- 10 GPT-5.6 Sol ↓5 68.4
Rule score (Sport snow)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 78.9
- 2 GPT-6 Astra ↓1 78.1
- 3
Gemini 4 Argon 73.6
- 4 Claude Opus 5.5 73.6
- 5
DeepSeek V4.1 Flash ↑1 73.2
- 6 GLM-5.3-Flash ↑3 72.8
- 7 GPT-5.6 Sol ↓2 72.0
- 8
Muse Spark 1.3 ↑3 71.8
- 9 Kimi K3 ↓2 71.4
- 10
Doubao Seed 2.1 Pro 68.7
Gemini 4 judge (Sport snow)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 86.2
- 2 GPT-6 Astra ↓1 85.9
- 3 Kimi K3 ↑4 81.0
- 4
DeepSeek V4.1 Flash ↑2 78.7
- 5 GPT-6.1 Sol ↓3 74.6
- 6 Claude Opus 5.5 ↓2 74.5
- 7 GPT-5.6 Sol ↓2 73.7
- 8
Doubao Seed 2.1 Pro ↑2 71.5
- 9
Muse Spark 1.3 ↑2 67.6
- 10
Gemini 3.8 Flash ↓2 66.2
Rule score (Sport motor)
11 models we track have a score here.
- 1 GPT-6 Astra 74.8
- 2 Claude Opus 5.5 ↑2 73.7
- 3 GPT-6.1 Sol ↓1 69.2
- 4 Kimi K3 ↑3 67.4
- 5
Gemini 3.8 Flash ↑3 66.4
- 6 GPT-5.6 Sol ↓1 65.8
- 7
DeepSeek V4.1 Flash ↓1 64.4
- 8 GLM-5.3-Flash ↑1 64.0
- 9
Gemini 4 Argon ↓6 60.7
- 10
Doubao Seed 2.1 Pro 55.0
Gemini 4 judge (Sport motor)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 81.9
- 2 GPT-6 Astra ↓1 81.8
- 3
DeepSeek V4.1 Flash ↑3 80.0
- 4 Claude Opus 5.5 79.4
- 5 GPT-5.6 Sol 79.3
- 6 Kimi K3 ↑1 74.7
- 7
Gemini 3.8 Flash ↑1 72.4
- 8
Gemini 4 Argon ↓5 70.1
- 9
Doubao Seed 2.1 Pro ↑1 60.0
- 10 GLM-5.3-Flash ↓1 57.2
Rule score (Sport water)
11 models we track have a score here.
- 1 GPT-6 Astra 85.7
- 2
Gemini 4 Argon ↑1 85.4
- 3 Kimi K3 ↑4 83.1
- 4
DeepSeek V4.1 Flash ↑2 82.9
- 5
Doubao Seed 2.1 Pro ↑5 82.3
- 6 Claude Opus 5.5 ↓2 82.1
- 7 GPT-5.6 Sol ↓2 80.8
- 8 GPT-6.1 Sol ↓6 80.5
- 9
Muse Spark 1.3 ↑2 79.5
- 10 GLM-5.3-Flash ↓1 78.3
Gemini 4 judge (Sport water)
11 models we track have a score here.
- 1 GPT-6 Astra 74.5
- 2 Kimi K3 ↑5 73.8
- 3 GPT-6.1 Sol ↓1 73.5
- 4
Gemini 4 Argon ↓1 71.6
- 5 GPT-5.6 Sol 69.3
- 6
DeepSeek V4.1 Flash 68.1
- 7
Gemini 3.8 Flash ↑1 64.7
- 8 GLM-5.3-Flash ↑1 63.4
- 9 Claude Opus 5.5 ↓5 60.4
- 10
Muse Spark 1.3 ↑1 58.8
Rule score (Sport athletics)
11 models we track have a score here.
- 1 GPT-6 Astra 83.3
- 2
Gemini 4 Argon ↑1 80.6
- 3 Claude Opus 5.5 ↑1 78.2
- 4 GPT-6.1 Sol ↓2 76.7
- 5
DeepSeek V4.1 Flash ↑1 76.0
- 6 GPT-5.6 Sol ↓1 75.7
- 7
Gemini 3.8 Flash ↑1 73.9
- 8 GLM-5.3-Flash ↑1 73.0
- 9
Muse Spark 1.3 ↑2 70.8
- 10 Kimi K3 ↓3 69.5
Gemini 4 judge (Sport athletics)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 81.7
- 2 GPT-6 Astra ↓1 80.0
- 3 GPT-6.1 Sol ↓1 80.0
- 4 Claude Opus 5.5 79.4
- 5 GPT-5.6 Sol 75.5
- 6
DeepSeek V4.1 Flash 72.0
- 7 Kimi K3 65.7
- 8
Gemini 3.8 Flash 64.2
- 9 GLM-5.3-Flash 62.5
- 10
Doubao Seed 2.1 Pro 59.2
Rule score (Sport animal)
11 models we track have a score here.
- 1 GPT-6 Astra 63.2
- 2
Gemini 4 Argon ↑1 60.4
- 3 GPT-6.1 Sol ↓1 60.3
- 4 Claude Opus 5.5 56.2
- 5 Kimi K3 ↑2 54.0
- 6 GLM-5.3-Flash ↑3 53.4
- 7
Gemini 3.8 Flash ↑1 51.9
- 8
DeepSeek V4.1 Flash ↓2 51.9
- 9 GPT-5.6 Sol ↓4 51.1
- 10
Muse Spark 1.3 ↑1 50.7
Gemini 4 judge (Sport animal)
11 models we track have a score here.
- 1 GPT-6 Astra 86.5
- 2 GPT-6.1 Sol 82.8
- 3 Claude Opus 5.5 ↑1 77.8
- 4
DeepSeek V4.1 Flash ↑2 75.9
- 5
Gemini 4 Argon ↓2 75.8
- 6 GPT-5.6 Sol ↓1 75.2
- 7
Doubao Seed 2.1 Pro ↑3 74.2
- 8 GLM-5.3-Flash ↑1 73.3
- 9 Kimi K3 ↓2 71.5
- 10
Gemini 3.8 Flash ↓2 70.0
Rule score (Sport obstacle)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 66.5
- 2 GPT-6 Astra ↓1 66.4
- 3 GPT-6.1 Sol ↓1 65.8
- 4
Gemini 4 Argon ↓1 65.1
- 5 Kimi K3 ↑2 60.5
- 6 GPT-5.6 Sol ↓1 59.3
- 7
DeepSeek V4.1 Flash ↓1 58.6
- 8 GLM-5.3-Flash ↑1 58.2
- 9
Gemini 3.8 Flash ↓1 57.5
- 10
Muse Spark 1.3 ↑1 54.9
Gemini 4 judge (Sport obstacle)
11 models we track have a score here.
- 1 GPT-6 Astra 91.0
- 2
Gemini 4 Argon ↑1 87.5
- 3 GPT-6.1 Sol ↓1 86.8
- 4 Claude Opus 5.5 86.0
- 5
DeepSeek V4.1 Flash ↑1 81.3
- 6 GPT-5.6 Sol ↓1 80.3
- 7 Kimi K3 77.7
- 8 GLM-5.3-Flash ↑1 75.5
- 9
Gemini 3.8 Flash ↓1 74.5
- 10
Doubao Seed 2.1 Pro 70.5
Rule score (Transit station)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 77.8
- 2
Gemini 4 Argon ↑1 75.1
- 3 GPT-6 Astra ↓2 75.0
- 4 Claude Opus 5.5 71.2
- 5 GPT-5.6 Sol 70.2
- 6 Kimi K3 ↑1 70.1
- 7
DeepSeek V4.1 Flash ↓1 69.9
- 8
Gemini 3.8 Flash 67.3
- 9 GLM-5.3-Flash 65.9
- 10
Muse Spark 1.3 ↑1 63.7
Gemini 4 judge (Transit station)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 82.0
- 2 GPT-6 Astra ↓1 81.2
- 3 Claude Opus 5.5 ↑1 77.4
- 4 GPT-6.1 Sol ↓2 76.5
- 5 GPT-5.6 Sol 75.8
- 6
DeepSeek V4.1 Flash 75.3
- 7 Kimi K3 73.6
- 8 GLM-5.3-Flash ↑1 61.8
- 9
Gemini 3.8 Flash ↓1 58.5
- 10
Doubao Seed 2.1 Pro 57.9
Rule score (Transit airport)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 89.9
- 2 GPT-6.1 Sol 89.2
- 3 GPT-6 Astra ↓2 88.4
- 4 Kimi K3 ↑3 87.8
- 5
Gemini 3.8 Flash ↑3 86.7
- 6 Claude Opus 5.5 ↓2 86.3
- 7 GPT-5.6 Sol ↓2 83.1
- 8
DeepSeek V4.1 Flash ↓2 83.0
- 9
Muse Spark 1.3 ↑2 79.9
- 10
Doubao Seed 2.1 Pro 75.8
Gemini 4 judge (Transit airport)
11 models we track have a score here.
- 1 GLM-5.3-Flash ↑8 80.1
- 2
Gemini 4 Argon ↑1 76.3
- 3 GPT-6.1 Sol ↓1 74.7
- 4 Claude Opus 5.5 74.0
- 5 GPT-5.6 Sol 69.6
- 6 Kimi K3 ↑1 65.3
- 7
DeepSeek V4.1 Flash ↓1 64.5
- 8 GPT-6 Astra ↓7 61.0
- 9
Gemini 3.8 Flash ↓1 58.7
- 10
Doubao Seed 2.1 Pro 54.2
Rule score (Transit port)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 82.2
- 2 GPT-6 Astra ↓1 81.3
- 3 Claude Opus 5.5 ↑1 79.0
- 4
Gemini 4 Argon ↓1 77.8
- 5 Kimi K3 ↑2 74.6
- 6 GPT-5.6 Sol ↓1 74.5
- 7 GLM-5.3-Flash ↑2 73.7
- 8
DeepSeek V4.1 Flash ↓2 72.0
- 9
Gemini 3.8 Flash ↓1 70.5
- 10
Doubao Seed 2.1 Pro 69.1
Gemini 4 judge (Transit port)
11 models we track have a score here.
- 1 GPT-6 Astra 75.3
- 2 GPT-6.1 Sol 72.3
- 3
Gemini 4 Argon 72.2
- 4 Claude Opus 5.5 69.7
- 5 GPT-5.6 Sol 69.0
- 6
DeepSeek V4.1 Flash 65.4
- 7 Kimi K3 65.1
- 8 GLM-5.3-Flash ↑1 62.1
- 9
Gemini 3.8 Flash ↓1 55.0
- 10
Doubao Seed 2.1 Pro 54.3
Rule score (Transit industrial)
11 models we track have a score here.
- 1 GPT-6 Astra 67.7
- 2 GPT-6.1 Sol 67.2
- 3
Gemini 4 Argon 64.4
- 4 Claude Opus 5.5 64.4
- 5 GPT-5.6 Sol 61.8
- 6
DeepSeek V4.1 Flash 60.9
- 7 Kimi K3 60.7
- 8
Gemini 3.8 Flash 58.4
- 9
Doubao Seed 2.1 Pro ↑1 57.1
- 10 GLM-5.3-Flash ↓1 55.3
Gemini 4 judge (Transit industrial)
11 models we track have a score here.
- 1 GPT-6 Astra 83.0
- 2 GPT-6.1 Sol 81.7
- 3 Claude Opus 5.5 ↑1 79.6
- 4
Gemini 4 Argon ↓1 77.4
- 5
DeepSeek V4.1 Flash ↑1 73.9
- 6 GPT-5.6 Sol ↓1 72.8
- 7 Kimi K3 69.8
- 8
Doubao Seed 2.1 Pro ↑2 64.5
- 9
Gemini 3.8 Flash ↓1 62.7
- 10 GLM-5.3-Flash ↓1 59.5
Rule score (Street city)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 54.9
- 2
Gemini 4 Argon ↑1 52.7
- 3 GPT-6.1 Sol ↓1 51.8
- 4 GPT-6 Astra ↓3 51.4
- 5 Kimi K3 ↑2 46.7
- 6 GPT-5.6 Sol ↓1 44.2
- 7
Gemini 3.8 Flash ↑1 43.3
- 8
DeepSeek V4.1 Flash ↓2 42.3
- 9
Muse Spark 1.3 ↑2 41.1
- 10 GLM-5.3-Flash ↓1 40.9
Gemini 4 judge (Street city)
11 models we track have a score here.
- 1 GPT-6 Astra 94.3
- 2
Gemini 4 Argon ↑1 89.2
- 3
DeepSeek V4.1 Flash ↑3 89.0
- 4 Claude Opus 5.5 88.8
- 5 GPT-6.1 Sol ↓3 86.5
- 6 GPT-5.6 Sol ↓1 84.5
- 7 GLM-5.3-Flash ↑2 78.6
- 8 Kimi K3 ↓1 75.6
- 9
Gemini 3.8 Flash ↓1 73.5
- 10
Doubao Seed 2.1 Pro 72.4
Rule score (Street residential)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 68.4
- 2 Claude Opus 5.5 ↑2 64.7
- 3 GPT-6 Astra ↓2 63.3
- 4 GPT-6.1 Sol ↓2 62.1
- 5 GPT-5.6 Sol 56.9
- 6
Gemini 3.8 Flash ↑2 56.4
- 7 Kimi K3 55.6
- 8 GLM-5.3-Flash ↑1 52.8
- 9
Doubao Seed 2.1 Pro ↑1 51.8
- 10
DeepSeek V4.1 Flash ↓4 51.8
Gemini 4 judge (Street residential)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 86.3
- 2 GPT-5.6 Sol ↑3 85.5
- 3 GPT-6.1 Sol ↓1 85.4
- 4
Gemini 4 Argon ↓1 85.4
- 5 GPT-6 Astra ↓4 84.6
- 6 Kimi K3 ↑1 80.1
- 7
DeepSeek V4.1 Flash ↓1 78.7
- 8 GLM-5.3-Flash ↑1 75.3
- 9
Gemini 3.8 Flash ↓1 73.8
- 10
Muse Spark 1.3 ↑1 72.7
Rule score (Street old town)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 61.4
- 2 GPT-6.1 Sol 59.5
- 3 GPT-6 Astra ↓2 59.0
- 4 GPT-5.6 Sol ↑1 56.3
- 5 Claude Opus 5.5 ↓1 54.3
- 6
DeepSeek V4.1 Flash 51.7
- 7 Kimi K3 51.6
- 8
Gemini 3.8 Flash 51.4
- 9 GLM-5.3-Flash 48.0
- 10
Doubao Seed 2.1 Pro 47.6
Gemini 4 judge (Street old town)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 90.5
- 2 GPT-6 Astra ↓1 89.0
- 3
Gemini 4 Argon 84.6
- 4 Claude Opus 5.5 81.1
- 5
DeepSeek V4.1 Flash ↑1 76.0
- 6 GPT-5.6 Sol ↓1 75.3
- 7 Kimi K3 69.7
- 8
Doubao Seed 2.1 Pro ↑2 67.1
- 9
Muse Spark 1.3 ↑2 63.9
- 10 GLM-5.3-Flash ↓1 62.7
Rule score (Street rooftop)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 56.1
- 2 GPT-6.1 Sol 53.0
- 3 Claude Opus 5.5 ↑1 53.0
- 4 GPT-6 Astra ↓3 52.2
- 5
Gemini 3.8 Flash ↑3 50.3
- 6 GPT-5.6 Sol ↓1 50.2
- 7
DeepSeek V4.1 Flash ↓1 43.9
- 8
Doubao Seed 2.1 Pro ↑2 43.5
- 9 GLM-5.3-Flash 43.0
- 10 Kimi K3 ↓3 40.6
Gemini 4 judge (Street rooftop)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 96.2
- 2
Gemini 4 Argon ↑1 95.4
- 3
DeepSeek V4.1 Flash ↑3 89.6
- 4 GPT-6 Astra ↓3 88.6
- 5 Claude Opus 5.5 ↓1 86.9
- 6 GPT-5.6 Sol ↓1 86.0
- 7
Gemini 3.8 Flash ↑1 83.9
- 8 Kimi K3 ↓1 77.8
- 9 GLM-5.3-Flash 71.6
- 10
Muse Spark 1.3 ↑1 66.0
Rule score (Nature snow)
11 models we track have a score here.
- 1 GPT-6 Astra 69.4
- 2 GPT-6.1 Sol 68.6
- 3
Gemini 4 Argon 63.4
- 4 Claude Opus 5.5 61.0
- 5 GPT-5.6 Sol 58.8
- 6
DeepSeek V4.1 Flash 58.2
- 7 Kimi K3 57.9
- 8 GLM-5.3-Flash ↑1 57.7
- 9
Gemini 3.8 Flash ↓1 56.5
- 10
Doubao Seed 2.1 Pro 54.5
Gemini 4 judge (Nature snow)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 83.8
- 2
Gemini 4 Argon ↑1 83.1
- 3 GPT-6 Astra ↓2 81.2
- 4 Claude Opus 5.5 77.6
- 5 GPT-5.6 Sol 74.8
- 6
DeepSeek V4.1 Flash 70.9
- 7
Gemini 3.8 Flash ↑1 70.6
- 8 Kimi K3 ↓1 67.3
- 9 GLM-5.3-Flash 67.1
- 10
Muse Spark 1.3 ↑1 67.0
Rule score (Nature mountain)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 70.1
- 2 GPT-6 Astra ↓1 69.2
- 3
Gemini 4 Argon 66.5
- 4 Claude Opus 5.5 66.0
- 5 GPT-5.6 Sol 64.3
- 6
Gemini 3.8 Flash ↑2 63.3
- 7 Kimi K3 60.9
- 8
DeepSeek V4.1 Flash ↓2 60.5
- 9
Muse Spark 1.3 ↑2 58.9
- 10 GLM-5.3-Flash ↓1 58.5
Gemini 4 judge (Nature mountain)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 84.0
- 2 GPT-6 Astra ↓1 82.9
- 3 GPT-6.1 Sol ↓1 81.1
- 4 GPT-5.6 Sol ↑1 74.0
- 5
DeepSeek V4.1 Flash ↑1 73.8
- 6 Claude Opus 5.5 ↓2 73.7
- 7 GLM-5.3-Flash ↑2 67.4
- 8 Kimi K3 ↓1 65.1
- 9
Gemini 3.8 Flash ↓1 63.5
- 10
Doubao Seed 2.1 Pro 61.3
Rule score (Nature water)
11 models we track have a score here.
- 1 GPT-6 Astra 78.2
- 2 Claude Opus 5.5 ↑2 77.5
- 3
Gemini 4 Argon 76.9
- 4 Kimi K3 ↑3 76.8
- 5 GPT-6.1 Sol ↓3 76.3
- 6
DeepSeek V4.1 Flash 71.5
- 7 GPT-5.6 Sol ↓2 71.4
- 8
Muse Spark 1.3 ↑3 70.0
- 9 GLM-5.3-Flash 69.3
- 10
Gemini 3.8 Flash ↓2 69.0
Gemini 4 judge (Nature water)
11 models we track have a score here.
- 1 GPT-6 Astra 81.3
- 2 Claude Opus 5.5 ↑2 80.9
- 3
Gemini 4 Argon 80.6
- 4 GPT-6.1 Sol ↓2 79.2
- 5
DeepSeek V4.1 Flash ↑1 75.4
- 6 GPT-5.6 Sol ↓1 73.8
- 7 Kimi K3 73.8
- 8 GLM-5.3-Flash ↑1 69.4
- 9
Gemini 3.8 Flash ↓1 65.5
- 10
Muse Spark 1.3 ↑1 63.1
Rule score (Setting Indoor)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 59.3
- 2 GPT-6 Astra ↓1 58.5
- 3 Claude Opus 5.5 ↑1 57.0
- 4
Gemini 4 Argon ↓1 56.6
- 5 GPT-5.6 Sol 51.1
- 6
DeepSeek V4.1 Flash 47.1
- 7 Kimi K3 46.4
- 8
Muse Spark 1.3 ↑3 46.0
- 9
Gemini 3.8 Flash ↓1 45.9
- 10 GLM-5.3-Flash ↓1 45.0
Gemini 4 judge (Setting Indoor)
11 models we track have a score here.
- 1 GPT-6 Astra 89.1
- 2 GPT-6.1 Sol 85.2
- 3
Gemini 4 Argon 83.4
- 4 Claude Opus 5.5 79.0
- 5
DeepSeek V4.1 Flash ↑1 77.3
- 6 GLM-5.3-Flash ↑3 75.1
- 7 GPT-5.6 Sol ↓2 74.6
- 8 Kimi K3 ↓1 71.3
- 9
Doubao Seed 2.1 Pro ↑1 70.5
- 10
Gemini 3.8 Flash ↓2 68.0
Setting Indoor
11 models we track have a score here.
- 1 GPT-6 Astra 73.8
- 2 GPT-6.1 Sol 72.3
- 3
Gemini 4 Argon 70.0
- 4 Claude Opus 5.5 68.0
- 5 GPT-5.6 Sol 62.9
- 6
DeepSeek V4.1 Flash 62.2
- 7 GLM-5.3-Flash ↑2 60.0
- 8 Kimi K3 ↓1 58.8
- 9
Doubao Seed 2.1 Pro ↑1 57.2
- 10
Gemini 3.8 Flash ↓2 57.0
Rule score (Setting Outdoor)
11 models we track have a score here.
- 1 GPT-6 Astra 56.8
- 2
Gemini 4 Argon ↑1 56.7
- 3 GPT-6.1 Sol ↓1 55.1
- 4 Claude Opus 5.5 53.3
- 5 GPT-5.6 Sol 48.9
- 6
Gemini 3.8 Flash ↑2 46.8
- 7
DeepSeek V4.1 Flash ↓1 45.2
- 8 Kimi K3 ↓1 45.2
- 9 GLM-5.3-Flash 42.8
- 10
Doubao Seed 2.1 Pro 41.1
Gemini 4 judge (Setting Outdoor)
11 models we track have a score here.
- 1 GPT-6 Astra 90.7
- 2 GPT-6.1 Sol 88.8
- 3
Gemini 4 Argon 88.1
- 4 GPT-5.6 Sol ↑1 83.4
- 5 Claude Opus 5.5 ↓1 82.9
- 6
DeepSeek V4.1 Flash 80.4
- 7 Kimi K3 74.2
- 8 GLM-5.3-Flash ↑1 70.8
- 9
Gemini 3.8 Flash ↓1 69.9
- 10
Doubao Seed 2.1 Pro 66.7
Setting Outdoor
11 models we track have a score here.
- 1 GPT-6 Astra 73.7
- 2
Gemini 4 Argon ↑1 72.4
- 3 GPT-6.1 Sol ↓1 72.0
- 4 Claude Opus 5.5 68.1
- 5 GPT-5.6 Sol 66.2
- 6
DeepSeek V4.1 Flash 62.8
- 7 Kimi K3 59.7
- 8
Gemini 3.8 Flash 58.3
- 9 GLM-5.3-Flash 56.8
- 10
Doubao Seed 2.1 Pro 53.9
Rule score (Setting Indoor and outdoor)
11 models we track have a score here.
- 1 GPT-6 Astra 54.8
- 2 GPT-6.1 Sol 53.5
- 3
Gemini 4 Argon 52.6
- 4 Claude Opus 5.5 51.2
- 5 GPT-5.6 Sol 46.5
- 6
DeepSeek V4.1 Flash 45.3
- 7
Gemini 3.8 Flash ↑1 45.0
- 8 Kimi K3 ↓1 43.7
- 9
Doubao Seed 2.1 Pro ↑1 42.9
- 10 GLM-5.3-Flash ↓1 39.1
Gemini 4 judge (Setting Indoor and outdoor)
11 models we track have a score here.
- 1 GPT-6 Astra 88.6
- 2 GPT-6.1 Sol 86.5
- 3
Gemini 4 Argon 82.3
- 4 Claude Opus 5.5 81.9
- 5
DeepSeek V4.1 Flash ↑1 76.7
- 6 GPT-5.6 Sol ↓1 72.4
- 7 Kimi K3 69.9
- 8
Gemini 3.8 Flash 64.8
- 9
Doubao Seed 2.1 Pro ↑1 64.0
- 10 GLM-5.3-Flash ↓1 62.5
Setting Indoor and outdoor
11 models we track have a score here.
- 1 GPT-6 Astra 71.7
- 2 GPT-6.1 Sol 70.0
- 3
Gemini 4 Argon 67.5
- 4 Claude Opus 5.5 66.5
- 5
DeepSeek V4.1 Flash ↑1 61.0
- 6 GPT-5.6 Sol ↓1 59.5
- 7 Kimi K3 56.8
- 8
Gemini 3.8 Flash 54.9
- 9
Doubao Seed 2.1 Pro ↑1 53.4
- 10 GLM-5.3-Flash ↓1 50.8
Rule score
11 models we track have a score here.
- 1 GPT-6 Astra 70.0
- 2 GPT-6.1 Sol 69.0
- 3 Claude Opus 5.5 ↑1 68.5
- 4
Gemini 4 Argon ↓1 68.4
- 5 GPT-5.6 Sol 62.9
- 6 Kimi K3 ↑1 61.5
- 7
DeepSeek V4.1 Flash ↓1 61.3
- 8
Gemini 3.8 Flash 60.7
- 9 GLM-5.3-Flash 57.9
- 10
Muse Spark 1.3 ↑1 57.2
Gemini 4 judge
11 models we track have a score here.
- 1 GPT-6 Astra 85.4
- 2 GPT-6.1 Sol 83.1
- 3
Gemini 4 Argon 81.7
- 4 Claude Opus 5.5 79.8
- 5
DeepSeek V4.1 Flash ↑1 77.9
- 6 GPT-5.6 Sol ↓1 77.5
- 7 Kimi K3 73.5
- 8 GLM-5.3-Flash ↑1 69.5
- 9
Gemini 3.8 Flash ↓1 68.7
- 10
Doubao Seed 2.1 Pro 66.2
Entities (Text)
11 models we track have a score here.
- 1 GPT-6 Astra 96.6
- 2 Kimi K3 ↑5 95.9
- 3 GPT-6.1 Sol ↓1 95.8
- 4
Gemini 4 Argon ↓1 95.8
- 5
Muse Spark 1.3 ↑6 95.6
- 6
DeepSeek V4.1 Flash 95.6
- 7 Claude Opus 5.5 ↓3 95.5
- 8
Gemini 3.8 Flash 95.3
- 9
Doubao Seed 2.1 Pro ↑1 94.6
- 10 GLM-5.3-Flash ↓1 93.8
Metric layout (Text)
11 models we track have a score here.
- 1 GPT-6 Astra 82.5
- 2 GPT-6.1 Sol 81.2
- 3
Gemini 4 Argon 74.5
- 4 Claude Opus 5.5 72.8
- 5 GPT-5.6 Sol 71.4
- 6
Gemini 3.8 Flash ↑2 69.7
- 7
Muse Spark 1.3 ↑4 65.1
- 8 Kimi K3 ↓1 64.0
- 9
DeepSeek V4.1 Flash ↓3 59.5
- 10 GLM-5.3-Flash ↓1 57.6
Routes (Text)
11 models we track have a score here.
- 1 GPT-6 Astra 83.2
- 2 GPT-6.1 Sol 81.6
- 3 Kimi K3 ↑4 81.1
- 4 Claude Opus 5.5 79.0
- 5
Gemini 4 Argon ↓2 76.8
- 6 GPT-5.6 Sol ↓1 75.5
- 7
DeepSeek V4.1 Flash ↓1 73.6
- 8 GLM-5.3-Flash ↑1 71.9
- 9
Muse Spark 1.3 ↑2 69.7
- 10
Doubao Seed 2.1 Pro 61.9
Events (Text)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 84.3
- 2
DeepSeek V4.1 Flash ↑4 84.0
- 3 Kimi K3 ↑4 83.2
- 4 Claude Opus 5.5 83.0
- 5 GPT-6 Astra ↓4 81.2
- 6 GPT-5.6 Sol ↓1 81.0
- 7
Gemini 3.8 Flash ↑1 80.0
- 8 GLM-5.3-Flash ↑1 79.5
- 9 GPT-6.1 Sol ↓7 78.6
- 10
Muse Spark 1.3 ↑1 77.0
Camera (Text)
11 models we track have a score here.
- 1 GPT-6 Astra 94.2
- 2
Gemini 4 Argon ↑1 94.1
- 3 GPT-6.1 Sol ↓1 92.0
- 4
DeepSeek V4.1 Flash ↑2 91.9
- 5 GPT-5.6 Sol 91.2
- 6 Claude Opus 5.5 ↓2 90.4
- 7
Gemini 3.8 Flash ↑1 89.7
- 8 Kimi K3 ↓1 89.1
- 9
Doubao Seed 2.1 Pro ↑1 84.0
- 10 GLM-5.3-Flash ↓1 83.8
Cuts (Text)
11 models we track have a score here.
- 1 GPT-5.6 Sol ↑4 94.2
- 2
DeepSeek V4.1 Flash ↑4 94.2
- 3 Claude Opus 5.5 ↑1 94.2
- 4
Gemini 4 Argon ↓1 94.2
- 5
Gemini 3.8 Flash ↑3 94.0
- 6 GPT-6 Astra ↓5 93.6
- 7 GPT-6.1 Sol ↓5 93.6
- 8 Kimi K3 ↓1 93.1
- 9
Muse Spark 1.3 ↑2 93.0
- 10 GLM-5.3-Flash ↓1 92.6
Physical validity (Text)
11 models we track have a score here.
- 1 GPT-6 Astra 68.1
- 2 GPT-6.1 Sol 68.0
- 3
Gemini 4 Argon 67.1
- 4 Claude Opus 5.5 65.9
- 5
Gemini 3.8 Flash ↑3 64.5
- 6 Kimi K3 ↑1 62.8
- 7
DeepSeek V4.1 Flash ↓1 62.0
- 8
Muse Spark 1.3 ↑3 59.9
- 9 GPT-5.6 Sol ↓4 59.0
- 10 GLM-5.3-Flash ↓1 58.5
Entities (Visual)
11 models we track have a score here.
- 1 Kimi K3 ↑6 81.7
- 2 Claude Opus 5.5 ↑2 81.7
- 3
Gemini 4 Argon 81.6
- 4 GPT-6 Astra ↓3 81.4
- 5 GPT-6.1 Sol ↓3 80.9
- 6
DeepSeek V4.1 Flash 79.7
- 7
Doubao Seed 2.1 Pro ↑3 79.4
- 8 GLM-5.3-Flash ↑1 79.2
- 9
Muse Spark 1.3 ↑2 79.1
- 10 GPT-5.6 Sol ↓5 77.9
Metric layout (Visual)
11 models we track have a score here.
- 1 GPT-6 Astra 36.1
- 2 GPT-6.1 Sol 33.9
- 3
Gemini 4 Argon 33.7
- 4 Claude Opus 5.5 27.3
- 5 Kimi K3 ↑2 22.0
- 6 GPT-5.6 Sol ↓1 20.7
- 7
DeepSeek V4.1 Flash ↓1 20.6
- 8
Gemini 3.8 Flash 18.2
- 9 GLM-5.3-Flash 17.1
- 10
Doubao Seed 2.1 Pro 15.5
Spaces (Visual)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 77.0
- 2 GPT-6.1 Sol 75.8
- 3 Claude Opus 5.5 ↑1 75.5
- 4 GPT-6 Astra ↓3 73.9
- 5
DeepSeek V4.1 Flash ↑1 68.0
- 6 GPT-5.6 Sol ↓1 66.4
- 7
Gemini 3.8 Flash ↑1 64.7
- 8 Kimi K3 ↓1 63.5
- 9 GLM-5.3-Flash 63.5
- 10
Doubao Seed 2.1 Pro 59.4
Picture structure (Visual)
11 models we track have a score here.
- 1 GPT-6 Astra 43.4
- 2 GPT-6.1 Sol 41.6
- 3
Gemini 4 Argon 40.5
- 4 Claude Opus 5.5 40.1
- 5 GPT-5.6 Sol 25.2
- 6
DeepSeek V4.1 Flash 25.1
- 7
Gemini 3.8 Flash ↑1 22.5
- 8 Kimi K3 ↓1 22.4
- 9 GLM-5.3-Flash 19.3
- 10
Doubao Seed 2.1 Pro 17.7
Routes (Visual)
11 models we track have a score here.
- 1 GPT-6 Astra 31.0
- 2 GPT-6.1 Sol 29.6
- 3
Gemini 4 Argon 28.4
- 4 Claude Opus 5.5 25.3
- 5 Kimi K3 ↑2 24.2
- 6 GPT-5.6 Sol ↓1 22.1
- 7
DeepSeek V4.1 Flash ↓1 21.9
- 8
Gemini 3.8 Flash 21.4
- 9
Muse Spark 1.3 ↑2 21.2
- 10
Doubao Seed 2.1 Pro 20.7
Events (Visual)
11 models we track have a score here.
- 1 GPT-6 Astra 61.8
- 2 GPT-6.1 Sol 60.9
- 3 Claude Opus 5.5 ↑1 59.2
- 4
Gemini 4 Argon ↓1 58.8
- 5 GPT-5.6 Sol 56.8
- 6 Kimi K3 ↑1 55.7
- 7 GLM-5.3-Flash ↑2 53.7
- 8
Gemini 3.8 Flash 49.0
- 9
DeepSeek V4.1 Flash ↓3 48.9
- 10
Doubao Seed 2.1 Pro 48.0
Camera (Visual)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 62.9
- 2 GPT-6 Astra ↓1 61.8
- 3 GPT-6.1 Sol ↓1 60.8
- 4 Claude Opus 5.5 58.3
- 5 GPT-5.6 Sol 50.8
- 6
DeepSeek V4.1 Flash 48.9
- 7 Kimi K3 48.8
- 8
Gemini 3.8 Flash 48.3
- 9
Doubao Seed 2.1 Pro ↑1 46.3
- 10 GLM-5.3-Flash ↓1 46.3
Cuts (Visual)
11 models we track have a score here.
- 1 GPT-6 Astra 85.9
- 2
Gemini 4 Argon ↑1 85.5
- 3 GPT-6.1 Sol ↓1 85.0
- 4
Gemini 3.8 Flash ↑4 83.5
- 5 Claude Opus 5.5 ↓1 83.2
- 6 GPT-5.6 Sol ↓1 82.4
- 7
DeepSeek V4.1 Flash ↓1 79.7
- 8
Muse Spark 1.3 ↑3 79.0
- 9 Kimi K3 ↓2 76.2
- 10
Doubao Seed 2.1 Pro 75.0
Physical validity (Visual)
11 models we track have a score here.
- 1 Claude Opus 5.5 ↑3 62.5
- 2
Gemini 4 Argon ↑1 56.2
- 3 GPT-6 Astra ↓2 54.0
- 4 GPT-6.1 Sol ↓2 53.4
- 5 GPT-5.6 Sol 49.8
- 6
Gemini 3.8 Flash ↑2 47.2
- 7 Kimi K3 45.1
- 8
DeepSeek V4.1 Flash ↓2 42.9
- 9
Muse Spark 1.3 ↑2 41.8
- 10 GLM-5.3-Flash ↓1 41.2
Rule score (People 1 to 2)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 55.4
- 2 GPT-6 Astra ↓1 55.2
- 3 Claude Opus 5.5 ↑1 55.1
- 4
Gemini 4 Argon ↓1 54.0
- 5 GPT-5.6 Sol 50.4
- 6 Kimi K3 ↑1 48.3
- 7
Gemini 3.8 Flash ↑1 46.7
- 8
DeepSeek V4.1 Flash ↓2 46.5
- 9 GLM-5.3-Flash 44.0
- 10
Doubao Seed 2.1 Pro 42.6
Gemini 4 judge (People 1 to 2)
11 models we track have a score here.
- 1 GPT-6 Astra 88.6
- 2
Gemini 4 Argon ↑1 86.3
- 3 GPT-6.1 Sol ↓1 85.0
- 4 GPT-5.6 Sol ↑1 80.8
- 5 Kimi K3 ↑2 79.7
- 6
DeepSeek V4.1 Flash 79.6
- 7 Claude Opus 5.5 ↓3 77.7
- 8 GLM-5.3-Flash ↑1 72.7
- 9
Doubao Seed 2.1 Pro ↑1 70.3
- 10
Gemini 3.8 Flash ↓2 69.4
People 1 to 2
11 models we track have a score here.
- 1 GPT-6 Astra 71.9
- 2 GPT-6.1 Sol 70.2
- 3
Gemini 4 Argon 70.1
- 4 Claude Opus 5.5 66.4
- 5 GPT-5.6 Sol 65.6
- 6 Kimi K3 ↑1 64.0
- 7
DeepSeek V4.1 Flash ↓1 63.1
- 8 GLM-5.3-Flash ↑1 58.3
- 9
Gemini 3.8 Flash ↓1 58.0
- 10
Doubao Seed 2.1 Pro 56.5
Rule score (People 3 to 5)
11 models we track have a score here.
- 1 GPT-6 Astra 57.6
- 2 GPT-6.1 Sol 56.5
- 3
Gemini 4 Argon 55.7
- 4 Claude Opus 5.5 52.8
- 5 GPT-5.6 Sol 49.0
- 6
Gemini 3.8 Flash ↑2 45.9
- 7
DeepSeek V4.1 Flash ↓1 45.6
- 8 Kimi K3 ↓1 44.5
- 9
Doubao Seed 2.1 Pro ↑1 42.9
- 10 GLM-5.3-Flash ↓1 41.6
Gemini 4 judge (People 3 to 5)
11 models we track have a score here.
- 1 GPT-6 Astra 89.7
- 2 GPT-6.1 Sol 87.2
- 3
Gemini 4 Argon 85.0
- 4 Claude Opus 5.5 82.7
- 5 GPT-5.6 Sol 79.0
- 6
DeepSeek V4.1 Flash 78.7
- 7 Kimi K3 72.0
- 8 GLM-5.3-Flash ↑1 69.4
- 9
Gemini 3.8 Flash ↓1 68.8
- 10
Doubao Seed 2.1 Pro 67.3
People 3 to 5
11 models we track have a score here.
- 1 GPT-6 Astra 73.6
- 2 GPT-6.1 Sol 71.9
- 3
Gemini 4 Argon 70.4
- 4 Claude Opus 5.5 67.7
- 5 GPT-5.6 Sol 64.0
- 6
DeepSeek V4.1 Flash 62.2
- 7 Kimi K3 58.3
- 8
Gemini 3.8 Flash 57.4
- 9 GLM-5.3-Flash 55.5
- 10
Doubao Seed 2.1 Pro 55.1
Rule score (People 6+)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 58.1
- 2 GPT-6 Astra ↓1 57.3
- 3 GPT-6.1 Sol ↓1 55.5
- 4 Claude Opus 5.5 54.4
- 5 GPT-5.6 Sol 47.8
- 6
Gemini 3.8 Flash ↑2 46.1
- 7
DeepSeek V4.1 Flash ↓1 45.1
- 8 GLM-5.3-Flash ↑1 43.0
- 9 Kimi K3 ↓2 42.9
- 10
Doubao Seed 2.1 Pro 40.8
Gemini 4 judge (People 6+)
11 models we track have a score here.
- 1 GPT-6 Astra 91.4
- 2 GPT-6.1 Sol 90.0
- 3
Gemini 4 Argon 85.7
- 4 Claude Opus 5.5 84.1
- 5
DeepSeek V4.1 Flash ↑1 77.8
- 6 GPT-5.6 Sol ↓1 75.8
- 7 GLM-5.3-Flash ↑2 69.0
- 8
Gemini 3.8 Flash 66.2
- 9 Kimi K3 ↓2 65.3
- 10
Doubao Seed 2.1 Pro 63.4
People 6+
11 models we track have a score here.
- 1 GPT-6 Astra 74.3
- 2 GPT-6.1 Sol 72.8
- 3
Gemini 4 Argon 71.9
- 4 Claude Opus 5.5 69.2
- 5 GPT-5.6 Sol 61.8
- 6
DeepSeek V4.1 Flash 61.5
- 7
Gemini 3.8 Flash ↑1 56.1
- 8 GLM-5.3-Flash ↑1 56.0
- 9 Kimi K3 ↓2 54.1
- 10
Doubao Seed 2.1 Pro 52.1
Rule score (Places 1)
11 models we track have a score here.
- 1 GPT-6 Astra 59.9
- 2 GPT-6.1 Sol 58.6
- 3
Gemini 4 Argon 58.4
- 4 Claude Opus 5.5 57.9
- 5 GPT-5.6 Sol 51.6
- 6
Gemini 3.8 Flash ↑2 47.3
- 7 Kimi K3 46.3
- 8
DeepSeek V4.1 Flash ↓2 46.0
- 9
Muse Spark 1.3 ↑2 45.7
- 10 GLM-5.3-Flash ↓1 44.8
Gemini 4 judge (Places 1)
11 models we track have a score here.
- 1 GPT-6 Astra 90.3
- 2
Gemini 4 Argon ↑1 87.3
- 3 GPT-6.1 Sol ↓1 85.8
- 4 GPT-5.6 Sol ↑1 81.7
- 5 Claude Opus 5.5 ↓1 81.6
- 6
DeepSeek V4.1 Flash 80.0
- 7 GLM-5.3-Flash ↑2 74.5
- 8 Kimi K3 ↓1 71.9
- 9
Gemini 3.8 Flash ↓1 71.3
- 10
Doubao Seed 2.1 Pro 71.0
Places 1
11 models we track have a score here.
- 1 GPT-6 Astra 75.1
- 2
Gemini 4 Argon ↑1 72.8
- 3 GPT-6.1 Sol ↓1 72.2
- 4 Claude Opus 5.5 69.7
- 5 GPT-5.6 Sol 66.6
- 6
DeepSeek V4.1 Flash 63.0
- 7 GLM-5.3-Flash ↑2 59.6
- 8
Gemini 3.8 Flash 59.3
- 9 Kimi K3 ↓2 59.1
- 10
Muse Spark 1.3 ↑1 57.5
Rule score (Places 2 to 3)
11 models we track have a score here.
- 1 GPT-6 Astra 57.3
- 2 GPT-6.1 Sol 55.6
- 3
Gemini 4 Argon 55.3
- 4 Claude Opus 5.5 52.8
- 5 GPT-5.6 Sol 48.3
- 6
Gemini 3.8 Flash ↑2 46.5
- 7
DeepSeek V4.1 Flash ↓1 45.8
- 8 Kimi K3 ↓1 45.0
- 9 GLM-5.3-Flash 43.3
- 10
Doubao Seed 2.1 Pro 42.8
Gemini 4 judge (Places 2 to 3)
11 models we track have a score here.
- 1 GPT-6 Astra 90.0
- 2 GPT-6.1 Sol 88.2
- 3
Gemini 4 Argon 83.7
- 4 Claude Opus 5.5 81.4
- 5 GPT-5.6 Sol 76.1
- 6
DeepSeek V4.1 Flash 75.5
- 7 Kimi K3 72.6
- 8
Gemini 3.8 Flash 69.6
- 9 GLM-5.3-Flash 66.4
- 10
Doubao Seed 2.1 Pro 64.3
Places 2 to 3
11 models we track have a score here.
- 1 GPT-6 Astra 73.7
- 2 GPT-6.1 Sol 71.9
- 3
Gemini 4 Argon 69.5
- 4 Claude Opus 5.5 67.1
- 5 GPT-5.6 Sol 62.2
- 6
DeepSeek V4.1 Flash 60.6
- 7 Kimi K3 58.8
- 8
Gemini 3.8 Flash 58.1
- 9 GLM-5.3-Flash 54.8
- 10
Doubao Seed 2.1 Pro 53.5
Rule score (Places 4+)
11 models we track have a score here.
- 1
Gemini 4 Argon ↑2 52.6
- 2 GPT-6.1 Sol 52.3
- 3 GPT-6 Astra ↓2 51.6
- 4 Claude Opus 5.5 49.3
- 5 GPT-5.6 Sol 46.1
- 6
DeepSeek V4.1 Flash 45.4
- 7
Gemini 3.8 Flash ↑1 44.0
- 8 Kimi K3 ↓1 43.8
- 9
Doubao Seed 2.1 Pro ↑1 40.3
- 10 GLM-5.3-Flash ↓1 38.5
Gemini 4 judge (Places 4+)
11 models we track have a score here.
- 1 GPT-6 Astra 88.8
- 2 GPT-6.1 Sol 88.5
- 3
Gemini 4 Argon 85.5
- 4 Claude Opus 5.5 81.9
- 5
DeepSeek V4.1 Flash ↑1 81.3
- 6 GPT-5.6 Sol ↓1 77.5
- 7 Kimi K3 73.2
- 8 GLM-5.3-Flash ↑1 68.9
- 9
Doubao Seed 2.1 Pro ↑1 65.2
- 10
Gemini 3.8 Flash ↓2 61.8
Places 4+
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 70.4
- 2 GPT-6 Astra ↓1 70.2
- 3
Gemini 4 Argon 69.1
- 4 Claude Opus 5.5 65.6
- 5
DeepSeek V4.1 Flash ↑1 63.3
- 6 GPT-5.6 Sol ↓1 61.8
- 7 Kimi K3 58.5
- 8 GLM-5.3-Flash ↑1 53.7
- 9
Gemini 3.8 Flash ↓1 52.9
- 10
Doubao Seed 2.1 Pro 52.7
Rule score (Shots 1 to 6)
11 models we track have a score here.
- 1 GPT-6 Astra 61.2
- 2 GPT-6.1 Sol 60.6
- 3
Gemini 4 Argon 60.3
- 4 Claude Opus 5.5 58.4
- 5 GPT-5.6 Sol 53.4
- 6
Gemini 3.8 Flash ↑2 50.2
- 7
DeepSeek V4.1 Flash ↓1 47.2
- 8 Kimi K3 ↓1 46.7
- 9
Muse Spark 1.3 ↑2 45.2
- 10
Doubao Seed 2.1 Pro 44.8
Gemini 4 judge (Shots 1 to 6)
11 models we track have a score here.
- 1 GPT-6 Astra 90.1
- 2 GPT-6.1 Sol 87.0
- 3
Gemini 4 Argon 86.9
- 4 Claude Opus 5.5 84.8
- 5
DeepSeek V4.1 Flash ↑1 82.1
- 6 GPT-5.6 Sol ↓1 81.3
- 7 GLM-5.3-Flash ↑2 75.2
- 8 Kimi K3 ↓1 73.0
- 9
Gemini 3.8 Flash ↓1 71.8
- 10
Muse Spark 1.3 ↑1 66.9
Shots 1 to 6
11 models we track have a score here.
- 1 GPT-6 Astra 75.6
- 2 GPT-6.1 Sol 73.8
- 3
Gemini 4 Argon 73.6
- 4 Claude Opus 5.5 71.6
- 5 GPT-5.6 Sol 67.3
- 6
DeepSeek V4.1 Flash 64.7
- 7
Gemini 3.8 Flash ↑1 61.0
- 8 Kimi K3 ↓1 59.9
- 9 GLM-5.3-Flash 59.6
- 10
Muse Spark 1.3 ↑1 56.1
Rule score (Shots 7 to 15)
11 models we track have a score here.
- 1 GPT-6 Astra 57.3
- 2
Gemini 4 Argon ↑1 55.3
- 3 GPT-6.1 Sol ↓1 54.9
- 4 Claude Opus 5.5 53.1
- 5 GPT-5.6 Sol 47.9
- 6 Kimi K3 ↑1 47.1
- 7
Gemini 3.8 Flash ↑1 44.8
- 8
DeepSeek V4.1 Flash ↓2 44.8
- 9 GLM-5.3-Flash 43.4
- 10
Doubao Seed 2.1 Pro 42.5
Gemini 4 judge (Shots 7 to 15)
11 models we track have a score here.
- 1 GPT-6 Astra 90.8
- 2 GPT-6.1 Sol 88.2
- 3
Gemini 4 Argon 86.0
- 4 Claude Opus 5.5 80.3
- 5 GPT-5.6 Sol 79.9
- 6
DeepSeek V4.1 Flash 76.6
- 7 Kimi K3 72.3
- 8 GLM-5.3-Flash ↑1 71.8
- 9
Gemini 3.8 Flash ↓1 69.4
- 10
Doubao Seed 2.1 Pro 69.1
Shots 7 to 15
11 models we track have a score here.
- 1 GPT-6 Astra 74.0
- 2 GPT-6.1 Sol 71.6
- 3
Gemini 4 Argon 70.7
- 4 Claude Opus 5.5 66.7
- 5 GPT-5.6 Sol 63.9
- 6
DeepSeek V4.1 Flash 60.7
- 7 Kimi K3 59.7
- 8 GLM-5.3-Flash ↑1 57.6
- 9
Gemini 3.8 Flash ↓1 57.1
- 10
Doubao Seed 2.1 Pro 55.8
Rule score (Shots 16+)
11 models we track have a score here.
- 1 GPT-6.1 Sol ↑1 53.0
- 2 GPT-6 Astra ↓1 52.7
- 3
Gemini 4 Argon 52.6
- 4 Claude Opus 5.5 50.8
- 5 GPT-5.6 Sol 46.5
- 6
DeepSeek V4.1 Flash 45.4
- 7
Gemini 3.8 Flash ↑1 44.0
- 8 Kimi K3 ↓1 42.2
- 9 GLM-5.3-Flash 40.8
- 10
Doubao Seed 2.1 Pro 39.8
Gemini 4 judge (Shots 16+)
11 models we track have a score here.
- 1 GPT-6 Astra 88.7
- 2 GPT-6.1 Sol 86.8
- 3
Gemini 4 Argon 84.0
- 4 Claude Opus 5.5 80.1
- 5
DeepSeek V4.1 Flash ↑1 78.0
- 6 GPT-5.6 Sol ↓1 75.3
- 7 Kimi K3 72.1
- 8
Doubao Seed 2.1 Pro ↑2 66.6
- 9 GLM-5.3-Flash 64.6
- 10
Gemini 3.8 Flash ↓2 64.3
Shots 16+
11 models we track have a score here.
- 1 GPT-6 Astra 70.7
- 2 GPT-6.1 Sol 69.9
- 3
Gemini 4 Argon 68.3
- 4 Claude Opus 5.5 65.5
- 5
DeepSeek V4.1 Flash ↑1 61.7
- 6 GPT-5.6 Sol ↓1 60.9
- 7 Kimi K3 57.1
- 8
Gemini 3.8 Flash 54.2
- 9
Doubao Seed 2.1 Pro ↑1 53.2
- 10 GLM-5.3-Flash ↓1 52.7
Rule score (Length < 30 s)
11 models we track have a score here.
- 1 GPT-6 Astra 60.7
- 2 GPT-6.1 Sol 59.3
- 3
Gemini 4 Argon 58.6
- 4 Claude Opus 5.5 58.5
- 5 GPT-5.6 Sol 53.8
- 6
Gemini 3.8 Flash ↑2 49.8
- 7 Kimi K3 49.1
- 8
DeepSeek V4.1 Flash ↓2 48.1
- 9 GLM-5.3-Flash 44.7
- 10
Doubao Seed 2.1 Pro 44.4
Gemini 4 judge (Length < 30 s)
11 models we track have a score here.
- 1 GPT-6 Astra 89.4
- 2
Gemini 4 Argon ↑1 87.7
- 3 GPT-6.1 Sol ↓1 86.4
- 4 Claude Opus 5.5 83.5
- 5
DeepSeek V4.1 Flash ↑1 81.7
- 6 GPT-5.6 Sol ↓1 80.5
- 7 Kimi K3 75.7
- 8 GLM-5.3-Flash ↑1 70.9
- 9
Gemini 3.8 Flash ↓1 70.3
- 10
Doubao Seed 2.1 Pro 68.6
Length < 30 s
11 models we track have a score here.
- 1 GPT-6 Astra 75.1
- 2
Gemini 4 Argon ↑1 73.1
- 3 GPT-6.1 Sol ↓1 72.8
- 4 Claude Opus 5.5 71.0
- 5 GPT-5.6 Sol 67.2
- 6
DeepSeek V4.1 Flash 64.9
- 7 Kimi K3 62.4
- 8
Gemini 3.8 Flash 60.0
- 9 GLM-5.3-Flash 57.8
- 10
Doubao Seed 2.1 Pro 56.5
Rule score (Length 30 to 44 s)
11 models we track have a score here.
- 1 GPT-6 Astra 57.0
- 2
Gemini 4 Argon ↑1 56.0
- 3 GPT-6.1 Sol ↓1 55.5
- 4 Claude Opus 5.5 50.5
- 5 GPT-5.6 Sol 48.0
- 6
Gemini 3.8 Flash ↑2 44.9
- 7
DeepSeek V4.1 Flash ↓1 44.8
- 8 Kimi K3 ↓1 44.3
- 9
Doubao Seed 2.1 Pro ↑1 43.0
- 10 GLM-5.3-Flash ↓1 42.8
Gemini 4 judge (Length 30 to 44 s)
11 models we track have a score here.
- 1 GPT-6 Astra 89.2
- 2 GPT-6.1 Sol 88.0
- 3
Gemini 4 Argon 85.7
- 4 GPT-5.6 Sol ↑1 80.1
- 5 Claude Opus 5.5 ↓1 79.7
- 6
DeepSeek V4.1 Flash 78.3
- 7 Kimi K3 73.1
- 8 GLM-5.3-Flash ↑1 71.3
- 9
Gemini 3.8 Flash ↓1 68.3
- 10
Doubao Seed 2.1 Pro 67.6
Length 30 to 44 s
11 models we track have a score here.
- 1 GPT-6 Astra 73.1
- 2 GPT-6.1 Sol 71.8
- 3
Gemini 4 Argon 70.9
- 4 Claude Opus 5.5 65.1
- 5 GPT-5.6 Sol 64.0
- 6
DeepSeek V4.1 Flash 61.5
- 7 Kimi K3 58.7
- 8 GLM-5.3-Flash ↑1 57.0
- 9
Gemini 3.8 Flash ↓1 56.6
- 10
Doubao Seed 2.1 Pro 55.3
Rule score (Length ≥ 45 s)
11 models we track have a score here.
- 1 GPT-6 Astra 54.3
- 2 GPT-6.1 Sol 54.1
- 3
Gemini 4 Argon 53.9
- 4 Claude Opus 5.5 53.2
- 5 GPT-5.6 Sol 46.6
- 6
DeepSeek V4.1 Flash 44.8
- 7
Gemini 3.8 Flash ↑1 44.7
- 8 Kimi K3 ↓1 43.3
- 9
Muse Spark 1.3 ↑2 41.3
- 10 GLM-5.3-Flash ↓1 41.2
Gemini 4 judge (Length ≥ 45 s)
11 models we track have a score here.
- 1 GPT-6 Astra 90.5
- 2 GPT-6.1 Sol 87.4
- 3
Gemini 4 Argon 84.1
- 4 Claude Opus 5.5 81.6
- 5
DeepSeek V4.1 Flash ↑1 77.2
- 6 GPT-5.6 Sol ↓1 76.5
- 7 Kimi K3 69.9
- 8 GLM-5.3-Flash ↑1 69.1
- 9
Gemini 3.8 Flash ↓1 67.0
- 10
Doubao Seed 2.1 Pro 65.9
Length ≥ 45 s
11 models we track have a score here.
- 1 GPT-6 Astra 72.4
- 2 GPT-6.1 Sol 70.7
- 3
Gemini 4 Argon 69.0
- 4 Claude Opus 5.5 67.4
- 5 GPT-5.6 Sol 61.6
- 6
DeepSeek V4.1 Flash 61.0
- 7 Kimi K3 56.6
- 8
Gemini 3.8 Flash 55.8
- 9 GLM-5.3-Flash 55.1
- 10
Doubao Seed 2.1 Pro 53.1
The best score, 77.7, costs $33 a task
Runs on the cost frontier score higher than any run that's cheaper. Gemini 4 Argon at reasoning: high; 4-hour task limit comes within 2.7 of the best for $4.45 a task, 7 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.
Source: Official VibePrevisBench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-6 Astra | reasoning: max; 4-hour task limit | Codex CLI | 77.7 | $33 | 148,594 |
| GPT-6.1 Sol | reasoning: max; 4-hour task limit | Codex CLI | 76.0 | $6.71 | 195,254 |
| reasoning: high; 4-hour task limit | Claude Code | 75.0 | $4.45 | 180,978 | |
| Claude Opus 5.5 | reasoning: high; 4-hour task limit | Claude Code | 74.1 | $7.82 | 137,249 |
| GPT-5.6 Sol | reasoning: max; 4-hour task limit | Codex CLI | 70.2 | $6.19 | 54,309 |
| reasoning: max; 4-hour task limit; model snapshot: 1017 | Claude Code | 69.6 | $1.89 | 260,679 | |
| Kimi K3 | reasoning: xhigh; 4-hour task limit | Claude Code | 67.5 | $18 | 226,367 |
| reasoning: high; 4-hour task limit | Claude Code | 64.7 | $2.73 | 166,256 | |
| GLM-5.3-Flash | reasoning: high; 4-hour task limit | Claude Code | 63.7 | $0.73 | 134,957 |
| reasoning: max; 4-hour task limit | Claude Code | 61.3 | $5.96 | 143,224 | |
| reasoning: xhigh; 4-hour task limit | Claude Code | 61.1 | $3.84 | 157,436 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official VibePrevisBench leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official VibePrevisBench leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Engineering and CAD
- EEBench In use 13 electrical-engineering circuit and PCB design tasks graded by deterministic build checks, simulation and BOM constraints. No LLM judge.
Gemini 4 Carbon 83.9%
- 3DCodeBench In use Agentic procedural 3D modeling via code: author engine-ready 3D assets through software APIs. Graded for executability and shape fidelity.
LongCat-2.5 55.4%
- CADGenBench Only in system cards Agents construct CAD models from design prompts. The generation split is graded for executability and geometric correctness. Claude Fable 5 45.1%
- PartBench Only in system cards Parametric mechanical part design and CAD-style generation, graded for geometric and manufacturing fidelity. Grok 4.6 58.5%
- Razavi-Bench Not in model cards Models explain what happens in analog circuits and why. A language model awards up to four points per answer against a reference and a rubric.
Gemini 4 Carbon 98.3%
- CircuitRubric Not in model cards Models write analog circuit netlists from design requests. A program compares transistor connections and relative device sizes with accepted reference circuits. Claude Opus 4.8 96%
Data audit
All 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (11 entries).
- Pass Models scored 11 models we track.
- Pass Newest score The newest is from September 2026, for Gemini 4 Argon.
- Pass Frontier models 4 of the top 10 have a score, led by GPT-6 Astra (#4).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on vibeprevisbench.github.io: 11 rows for models we track at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.