WordleArenas
Models guess a five-letter word from colored clues. Each arena changes how many clues give false feedback after a guess, from zero to five.
Liu reports 41 models across 3,749 games in the paper. The later batch results cover 52 configurations across 4,163 games with different trial counts.
- Made by
- Chang Liu
- Released
- February 2026
- Kind
- Puzzles and chess
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Wordle
- Items from
- Generated
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
-
Gemini 3.1 Pro #40 on the capability index
- Last model added
- April 2026
- Scores refreshed
- November 20th, 2026
2 models share the top score of 100% among models released by April 2026 2 models share the top score of 100% among models released by April 2026
Best score per entry 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. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in April 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 Astra
- and 16 more
- 100.0 Claude Opus 4
- 100.0 Claude Sonnet 4
- 95.0
DeepSeek-V4-Pro
- 62.2
DeepSeek-R1
- 57.2
DeepSeek-V4-Flash
- 51.7
Gemma 4 31B IT
- 42.0
Gemini 3 Flash
- 41.9 Glm 5 cloud
- 39.4 Kimi K2.5 cloud
- 38.3
Gemini 3.1 Pro
- 30.7 Qwen3 30b
- 25.3
Qwen3.5-122B-A10B
- 25.0
Gemini 2.5 Flash
- 23.2 Qwen3.5 cloud
- 21.7
Gemma 4 26B A4B
- 21.3
Qwen3-14B
- 19.7
Gemini 2.5 Pro
- 14.3
Qwen3-8B
- 13.9 o3
- 13.1 Claude Haiku 4.5
- 12.8
Qwen3-32B
- 11.7
Qwen3.5-27B
- 9.80 GPT-5 mini
- 9.50 Claude Sonnet 4.6
- 9.20 Deepseek Chat
- 8.80 GPT-4o
- 7.30
Gemma 3 27B
- 7.30 GPT-5
- 7.20 GPT-5.1
- 5.70 Llama3.1
- 5.60
Gemini 2.0 Flash
- 3.80 Llama4
- 3.50 GPT-4o mini
- 1.70
Gemma 4 E2B
- 1.30 Minimax M2.5 cloud
- 0.0 Codex Mini
- 0.0 Deepseek V2 16b
- 0.0 Llama3
- 0.0 Llama3.2
- 0.0 SmolLM3 3b
- 0.0
SmolLM2-1.7B
- 0.0
SmolLM2-135M
- 0.0 o4-mini
- 0.0
Qwen3.5-35B-A3B
- 0.0
Gemma 4 E4B
- 1 Claude Opus 4 100%
- 2 Claude Sonnet 4 100%
- 3
DeepSeek-V4-Pro 95%
- 4
DeepSeek-R1 62.2%
- 5
DeepSeek-V4-Flash 57.2%
- 6
Gemma 4 31B IT 51.7%
- 7
Gemini 3 Flash 42%
- 8 Glm 5 cloud 41.9%
- 9 Kimi K2.5 cloud 39.4%
- 10
Gemini 3.1 Pro 38.3%
- 11 Qwen3 30b 30.7%
- 12
Qwen3.5-122B-A10B 25.3%
- 13
Gemini 2.5 Flash Sep 2025 25%
- 14 Qwen3.5 cloud 23.2%
- 15
Gemma 4 26B A4B 21.7%
- 16
Qwen3-14B 21.3%
- 17
Gemini 2.5 Pro Jun 2025 19.7%
- 18
Qwen3-8B 14.3%
- 19 o3 13.9%
- 20 Claude Haiku 4.5 13.1%
- 21
Qwen3-32B 12.8%
- 22
Qwen3.5-27B 11.7%
- 23 GPT-5 mini 9.80%
- 24 Claude Sonnet 4.6 9.50%
- 25 Deepseek Chat 9.20%
- 26 GPT-4o Mar 2025 8.80%
- 27
Gemma 3 27B 7.30%
- 28 GPT-5 7.30%
- 29 GPT-5.1 7.20%
- 30 Llama3.1 5.70%
- 31
Gemini 2.0 Flash Feb 2025 5.60%
- 32 Llama4 3.80%
- 33 GPT-4o mini 3.50%
- 34
Gemma 4 E2B 1.70%
- 35 Minimax M2.5 cloud 1.30%
- 36 Codex Mini 0%
- 37 Deepseek V2 16b 0%
- 38 Llama3 0%
- 39 Llama3.2 0%
- 40 SmolLM3 3b 0%
- 41
SmolLM2-1.7B 0%
- 42
SmolLM2-135M 0%
- 43 o4-mini 0%
- 44
Qwen3.5-35B-A3B 0%
- 45
Gemma 4 E4B 0%
Source: Official WordleArenas leaderboard
- 100.0 Claude Opus 4
- 100.0 Claude Sonnet 4
- 95.0
DeepSeek-V4-Pro
- 62.2
DeepSeek-R1
- 57.2
DeepSeek-V4-Flash
- 51.7
Gemma 4 31B IT
- 42.0
Gemini 3 Flash
- 38.3
Gemini 3.1 Pro
- 25.3
Qwen3.5-122B-A10B
- 25.0
Gemini 2.5 Flash
- 21.7
Gemma 4 26B A4B
- 21.3
Qwen3-14B
- 19.7
Gemini 2.5 Pro
- 14.3
Qwen3-8B
- 13.9 o3
- 13.1 Claude Haiku 4.5
- 12.8
Qwen3-32B
- 11.7
Qwen3.5-27B
- 9.80 GPT-5 mini
- 9.50 Claude Sonnet 4.6
- 8.80 GPT-4o
- 7.30
Gemma 3 27B
- 7.30 GPT-5
- 7.20 GPT-5.1
- 5.60
Gemini 2.0 Flash
- 3.50 GPT-4o mini
- 1.70
Gemma 4 E2B
- 0.0
SmolLM2-1.7B
- 0.0
SmolLM2-135M
- 0.0 o4-mini
- 0.0
Qwen3.5-35B-A3B
- 0.0
Gemma 4 E4B
- 1 Claude Opus 4 100%
- 2 Claude Sonnet 4 100%
- 3
DeepSeek-V4-Pro 95%
- 4
DeepSeek-R1 62.2%
- 5
DeepSeek-V4-Flash 57.2%
- 6
Gemma 4 31B IT 51.7%
- 7
Gemini 3 Flash 42%
- 8
Gemini 3.1 Pro 38.3%
- 9
Qwen3.5-122B-A10B 25.3%
- 10
Gemini 2.5 Flash Sep 2025 25%
- 11
Gemma 4 26B A4B 21.7%
- 12
Qwen3-14B 21.3%
- 13
Gemini 2.5 Pro Jun 2025 19.7%
- 14
Qwen3-8B 14.3%
- 15 o3 13.9%
- 16 Claude Haiku 4.5 13.1%
- 17
Qwen3-32B 12.8%
- 18
Qwen3.5-27B 11.7%
- 19 GPT-5 mini 9.80%
- 20 Claude Sonnet 4.6 9.50%
- 21 GPT-4o Mar 2025 8.80%
- 22
Gemma 3 27B 7.30%
- 23 GPT-5 7.30%
- 24 GPT-5.1 7.20%
- 25
Gemini 2.0 Flash Feb 2025 5.60%
- 26 GPT-4o mini 3.50%
- 27
Gemma 4 E2B 1.70%
- 28
SmolLM2-1.7B 0%
- 29
SmolLM2-135M 0%
- 30 o4-mini 0%
- 31
Qwen3.5-35B-A3B 0%
- 32
Gemma 4 E4B 0%
Source: Official WordleArenas leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 4 | 100% | May 2025 | Official leaderboard | |
| 2 | Claude Sonnet 4 | 100% | May 2025 | Official leaderboard | |
| 3 | | 95% | Apr 2026 | Official leaderboard | |
| 4 | | 62.2% | Jan 2025 | Official leaderboard | |
| 5 | | 57.2% | Apr 2026 | Official leaderboard | |
| 6 | | 51.7% | Apr 2026 | Official leaderboard | |
| 7 | | 42% | Dec 2025 | Official leaderboard | |
| 8 | Glm 5 cloud | 41.9% | Official leaderboard | ||
| 9 | Kimi K2.5 cloud | 39.4% | Official leaderboard | ||
| 10 | | 38.3% | Feb 2026 | Official leaderboard | |
| 11 | Qwen3 30b | 30.7% | Official leaderboard | ||
| 12 | | 25.3% | Feb 2026 | Official leaderboard | |
| 13 | | 25% | Sep 2025 | Official leaderboard | |
| 14 | Qwen3.5 cloud | 23.2% | Official leaderboard | ||
| 15 | | 21.7% | Apr 2026 | Official leaderboard | |
| 16 | | 21.3% | Apr 2025 | Official leaderboard | |
| 17 | | 19.7% | Jun 2025 | Official leaderboard | |
| 18 | | 14.3% | Apr 2025 | Official leaderboard | |
| 19 | o3 | 13.9% | Apr 2025 | Official leaderboard | |
| 20 | Claude Haiku 4.5 | 13.1% | Oct 2025 | Official leaderboard | |
| 21 | | 12.8% | Apr 2025 | Official leaderboard | |
| 22 | | 11.7% | Feb 2026 | Official leaderboard | |
| 23 | GPT-5 mini | 9.80% | Aug 2025 | Official leaderboard | |
| 24 | Claude Sonnet 4.6 | 9.50% | Feb 2026 | Official leaderboard | |
| 25 | Deepseek Chat | 9.20% | Official leaderboard | ||
| 26 | GPT-4o Mar 2025 | 8.80% | Mar 2025 | Official leaderboard | |
| 27 | | 7.30% | Mar 2025 | Official leaderboard | |
| 28 | GPT-5 | 7.30% | Aug 2025 | Official leaderboard | |
| 29 | GPT-5.1 | 7.20% | Nov 2025 | Official leaderboard | |
| 30 | Llama3.1 | 5.70% | Official leaderboard | ||
| 31 | | 5.60% | Feb 2025 | Official leaderboard | |
| 32 | Llama4 | 3.80% | Official leaderboard | ||
| 33 | GPT-4o mini | 3.50% | Jul 2024 | Official leaderboard | |
| 34 | | 1.70% | Mar 2026 | Official leaderboard | |
| 35 | Minimax M2.5 cloud | 1.30% | Official leaderboard | ||
| 36 | Codex Mini | 0% | Official leaderboard | ||
| 37 | Deepseek V2 16b | 0% | Official leaderboard | ||
| 38 | Llama3 | 0% | Official leaderboard | ||
| 39 | Llama3.2 | 0% | Official leaderboard | ||
| 40 | SmolLM3 3b | 0% | Official leaderboard | ||
| 41 | | 0% | Oct 2024 | Official leaderboard | |
| 42 | | 0% | Oct 2024 | Official leaderboard | |
| 43 | o4-mini | 0% | Apr 2025 | Official leaderboard | |
| 44 | | 0% | Feb 2026 | Official leaderboard | |
| 45 | | 0% | Mar 2026 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 4 | 100% | May 2025 | Official leaderboard | |
| 2 | Claude Sonnet 4 | 100% | May 2025 | Official leaderboard | |
| 3 | | 95% | Apr 2026 | Official leaderboard | |
| 4 | | 62.2% | Jan 2025 | Official leaderboard | |
| 5 | | 57.2% | Apr 2026 | Official leaderboard | |
| 6 | | 51.7% | Apr 2026 | Official leaderboard | |
| 7 | | 42% | Dec 2025 | Official leaderboard | |
| 8 | | 38.3% | Feb 2026 | Official leaderboard | |
| 9 | | 25.3% | Feb 2026 | Official leaderboard | |
| 10 | | 25% | Sep 2025 | Official leaderboard | |
| 11 | | 21.7% | Apr 2026 | Official leaderboard | |
| 12 | | 21.3% | Apr 2025 | Official leaderboard | |
| 13 | | 19.7% | Jun 2025 | Official leaderboard | |
| 14 | | 14.3% | Apr 2025 | Official leaderboard | |
| 15 | o3 | 13.9% | Apr 2025 | Official leaderboard | |
| 16 | Claude Haiku 4.5 | 13.1% | Oct 2025 | Official leaderboard | |
| 17 | | 12.8% | Apr 2025 | Official leaderboard | |
| 18 | | 11.7% | Feb 2026 | Official leaderboard | |
| 19 | GPT-5 mini | 9.80% | Aug 2025 | Official leaderboard | |
| 20 | Claude Sonnet 4.6 | 9.50% | Feb 2026 | Official leaderboard | |
| 21 | GPT-4o Mar 2025 | 8.80% | Mar 2025 | Official leaderboard | |
| 22 | | 7.30% | Mar 2025 | Official leaderboard | |
| 23 | GPT-5 | 7.30% | Aug 2025 | Official leaderboard | |
| 24 | GPT-5.1 | 7.20% | Nov 2025 | Official leaderboard | |
| 25 | | 5.60% | Feb 2025 | Official leaderboard | |
| 26 | GPT-4o mini | 3.50% | Jul 2024 | Official leaderboard | |
| 27 | | 1.70% | Mar 2026 | Official leaderboard | |
| 28 | | 0% | Oct 2024 | Official leaderboard | |
| 29 | | 0% | Oct 2024 | Official leaderboard | |
| 30 | o4-mini | 0% | Apr 2025 | Official leaderboard | |
| 31 | | 0% | Feb 2026 | Official leaderboard | |
| 32 | | 0% | Mar 2026 | Official leaderboard |
Rankings by part
We read 6 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Wordle, Fibble and Fibble². DeepSeek-R1 leads 4 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Wordle
31 models we track have a score here.
- 1 Claude Opus 4 100%
- 2 Claude Sonnet 4 100%
- 3
Gemini 3 Flash ↑4 100%
- 4
Gemma 4 26B A4B ↑7 100%
- 5
Gemma 4 31B IT ↑1 100%
- 6
DeepSeek-V4-Pro ↓3 100%
- 7
Gemini 3.1 Pro ↑1 95%
- 8
Qwen3-14B ↑4 91.7%
- 9
DeepSeek-R1 ↓5 90%
- 10
Gemini 2.5 Pro Jun 2025 ↑3 90%
Fibble
29 models we track have a score here.
- 1
DeepSeek-R1 ↑3 100%
- 2
Gemma 4 31B IT ↑4 100%
- 3
DeepSeek-V4-Flash ↑2 100%
- 4
DeepSeek-V4-Pro ↓1 90%
- 5
Gemini 3.1 Pro ↑3 87.5%
- 6
Gemini 3 Flash ↑1 70%
- 7
Gemini 2.5 Flash Sep 2025 ↑3 50%
- 8
Qwen3.5-122B-A10B ↑1 42.1%
- 9 Claude Haiku 4.5 ↑7 30%
- 10
Qwen3.5-27B ↑8 30%
Fibble²
28 models we track have a score here.
- 1
DeepSeek-R1 ↑3 80%
- 2
DeepSeek-V4-Flash ↑3 60%
- 3
Gemma 4 31B IT ↑3 40%
- 4
Gemini 3.1 Pro ↑4 31.6%
- 5
Qwen3.5-122B-A10B ↑4 16.7%
- 6
Gemini 3 Flash ↑1 6.70%
- 7 GPT-4o mini ↑19 0%
- 8
SmolLM2-1.7B ↑20 0%
- 9
SmolLM2-135M ↑20 0%
- 10
Gemini 2.0 Flash Feb 2025 ↑15 0%
Fibble³
27 models we track have a score here.
- 1
Qwen3-32B ↑16 33.3%
- 2
Gemini 3 Flash ↑5 3.30%
- 3 GPT-4o mini ↑23 0%
- 4
SmolLM2-1.7B ↑24 0%
- 5
SmolLM2-135M ↑24 0%
- 6
DeepSeek-R1 ↓2 0%
- 7
Gemini 2.0 Flash Feb 2025 ↑18 0%
- 8
Gemma 3 27B ↑14 0%
- 9 GPT-4o Mar 2025 ↑12 0%
- 10 o3 ↑5 0%
Fibble⁴
26 models we track have a score here.
- 1
DeepSeek-R1 ↑3 33.3%
- 2
DeepSeek-V4-Flash ↑3 33.3%
- 3
Gemini 3.1 Pro ↑5 5.10%
- 4 GPT-4o mini ↑22 0%
- 5
SmolLM2-1.7B ↑23 0%
- 6
SmolLM2-135M ↑23 0%
- 7
Gemini 2.0 Flash Feb 2025 ↑18 0%
- 8
Gemma 3 27B ↑14 0%
- 9 GPT-4o Mar 2025 ↑12 0%
- 10 o3 ↑5 0%
Fibble⁵
27 models we track have a score here.
- 1
DeepSeek-R1 ↑3 70%
- 2
Gemma 4 31B IT ↑4 70%
- 3
DeepSeek-V4-Flash ↑2 60%
- 4
Gemini 3 Flash ↑3 30%
- 5
Gemma 4 26B A4B ↑6 30%
- 6
Qwen3-14B ↑6 27.3%
- 7
Qwen3-8B ↑7 27.3%
- 8
Qwen3.5-122B-A10B ↑1 18.2%
- 9
Gemini 3.1 Pro ↓1 10.3%
- 10
Gemma 3 27B ↑12 9.10%
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official WordleArenas leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Claude Opus 4 | 100% | ||||
| Claude Sonnet 4 | 100% | ||||
| 95% | |||||
| 62.2% | |||||
| 57.2% | |||||
| 51.7% | |||||
| 42% | |||||
| 38.3% | |||||
| 25.3% | |||||
| 25% | |||||
| 21.7% | |||||
| 21.3% | |||||
| 19.7% | |||||
| 14.3% | |||||
| o3 | 13.9% | ||||
| Claude Haiku 4.5 | 13.1% | ||||
| 12.8% | |||||
| 11.7% | |||||
| GPT-5 mini | 9.80% | ||||
| Claude Sonnet 4.6 | 9.50% | ||||
| GPT-4o Mar 2025 | 8.80% | ||||
| 7.30% | |||||
| GPT-5 | 7.30% | ||||
| no_think | 7.20% | ||||
| GPT-5.1 | 7.20% | ||||
| 5.60% | |||||
| no_think | 5.00% | ||||
| GPT-4o mini | 3.50% | ||||
| no_think | 3.30% | ||||
| no_think | 1.70% | ||||
| 1.70% | |||||
| no_think | 1.70% | ||||
| 0% | |||||
| 0% | |||||
| o4-mini | 0% | ||||
| no_think | 0% | ||||
| 0% | |||||
| 0% |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official WordleArenas leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Claude Opus 4 | 100% | ||||
| Claude Sonnet 4 | 100% | ||||
| 95% | |||||
| 62.2% | |||||
| 57.2% | |||||
| 51.7% | |||||
| 42% | |||||
| 38.3% | |||||
| 25.3% | |||||
| 25% | |||||
| 21.7% | |||||
| 21.3% | |||||
| 19.7% | |||||
| 14.3% | |||||
| o3 | 13.9% | ||||
| Claude Haiku 4.5 | 13.1% | ||||
| 12.8% | |||||
| 11.7% | |||||
| GPT-5 mini | 9.80% | ||||
| Claude Sonnet 4.6 | 9.50% | ||||
| GPT-4o Mar 2025 | 8.80% | ||||
| 7.30% | |||||
| GPT-5 | 7.30% | ||||
| no_think | 7.20% | ||||
| GPT-5.1 | 7.20% | ||||
| 5.60% | |||||
| no_think | 5.00% | ||||
| GPT-4o mini | 3.50% | ||||
| no_think | 3.30% | ||||
| no_think | 1.70% | ||||
| 1.70% | |||||
| no_think | 1.70% | ||||
| 0% | |||||
| 0% | |||||
| o4-mini | 0% | ||||
| no_think | 0% | ||||
| 0% | |||||
| 0% |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official WordleArenas leaderboard
Claude Opus 4 on the official leaderboard passed 85% in May 2025
Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official WordleArenas 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 Games and long tasks
- ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6.2 Astra 100%
- Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Each game scores in-game progress or outcomes, and the board ranks models by their average rank across the games.
Gemini 2.5 Pro 3.90
- MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks.
Seed2.0 Pro 49%
- TextQuests Only in system cards Agents play 25 classic interactive fiction games. The score is harmful actions per 100 steps, where lower is better. The board also reports game progress.
Llama 4 Scout 7.00
- GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. GPT-6.1 Terra 76.8%
- Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%
Data audit
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (45 entries).
- Pass Models scored 32 models we track, plus 13 other entries on its leaderboard.
- Note Newest score The newest came 7 months ago, in April 2026, for DeepSeek-V4-Flash. That’s expected for a saturated test.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Gemini 3.1 Pro (#40).
- Note Headroom Saturated. The best is 100%, past 85% since May 2025.
Leaderboard import 3 pass
- Pass Importer We read the board on drchangliu.github.io: 38 rows for models we track and 13 other entries 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.