Benchmark Almanac

OpenDesign Arena

Kimi K3.1 Pro 91.0 leads since October 2026
Not in model cards

Models create design artifacts in the OpenDesign tool, and the board reports an evaluation score out of 100 with cost and speed per artifact.

Vendor board inside a design product. How scores are judged needs a check.

Made by
OpenDesign
Kind
Web and app front ends
Level
Everyday
Scored
Graded by a model
Answers
Open-ended
Tasks
Kept private
Strongest model tested
GPT-6 Astra #3 on the capability index
Last model added
November 2026
Scores refreshed
November 12th, 2026
23
Entries on its own leaderboard, 21 models we track

Kimi K3.1 Pro leads with 91.0

Best score per entry from the benchmark's own leaderboard, read on November 12th, 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.

23 entries. Scroll sideways for the rest.

  1. 91.0 Kimi K3.1 Pro
  2. 90.3 GPT-6.1 Terra
  3. 90.1 GPT-6.1 Sol
  4. 90.0 Tencent Hy4
  5. 89.9 GPT-6 Sol
  6. 86.8 Claude Sonnet 5.5
  7. 83.5 MiMo 2.6 Flash
  8. 83.1 GPT-6 Luna
  9. 82.9 Grok 4.7
  10. 82.7 GPT-6 Astra
  11. 82.2 Claude Haiku 5.5
  12. 81.7 MiniMax-M3.1-Flash
  13. 81.2 DeepSeek V4.1 Flash
  14. 81.2 Gemini 4 Argon
  15. 80.3 Claude Fable 5.1
  16. 80.1 GLM-5.3 Flash-X
  17. 77.6 GPT-5.6 Sol
  18. 76.2 Muse Wonder
  19. 74.6 Tencent Hy4 preview
  20. 72.0 Qwen 3.8 Max
  21. 68.9 Gemini 3.8 Flash
  22. 66.6 Muse Spark 1.3
  23. 65.7 Kimi K3
  1. 1 Kimi K3.1 Pro 91.0
  2. 2 GPT-6.1 Terra 90.3
  3. 3 GPT-6.1 Sol 90.1
  4. 4 Tencent Hy4 90.0
  5. 5 GPT-6 Sol 89.9
  6. 6 Claude Sonnet 5.5 86.8
  7. 7 MiMo 2.6 Flash 83.5
  8. 8 GPT-6 Luna 83.1
  9. 9 Grok 4.7 82.9
  10. 10 GPT-6 Astra 82.7
  11. 11 Claude Haiku 5.5 82.2
  12. 12 MiniMax-M3.1-Flash 81.7
  13. 13 DeepSeek V4.1 Flash 81.2
  14. 14 Gemini 4 Argon 81.2
  15. 15 Claude Fable 5.1 80.3
  16. 16 GLM-5.3 Flash-X 80.1
  17. 17 GPT-5.6 Sol 77.6
  18. 18 Muse Wonder 76.2
  19. 19 Tencent Hy4 preview 74.6
  20. 20 Qwen 3.8 Max 72.0
  21. 21 Gemini 3.8 Flash 68.9
  22. 22 Muse Spark 1.3 66.6
  23. 23 Kimi K3 65.7

Source: Official OpenDesign Arena leaderboard

#EntryScoreSettingDateSource
1 Kimi K3.1 Pro 91.0 Oct 2026 Official leaderboard
2 GPT-6.1 Terra 90.3 Oct 2026 Official leaderboard
3 GPT-6.1 Sol 90.1 Sep 2026 Official leaderboard
4 Tencent Hy4 90.0 Nov 2026 Official leaderboard
5 GPT-6 Sol 89.9 Sep 2026 Official leaderboard
6 Claude Sonnet 5.5 86.8 Sep 2026 Official leaderboard
7 MiMo 2.6 Flash 83.5 Official leaderboard
8 GPT-6 Luna 83.1 Sep 2026 Official leaderboard
9 Grok 4.7 82.9 Sep 2026 Official leaderboard
10 GPT-6 Astra 82.7 Sep 2026 Official leaderboard
11 Claude Haiku 5.5 82.2 Oct 2026 Official leaderboard
12 MiniMax-M3.1-Flash 81.7 Oct 2026 Official leaderboard
13 DeepSeek V4.1 Flash 81.2 Sep 2026 Official leaderboard
14 Gemini 4 Argon 81.2 Sep 2026 Official leaderboard
15 Claude Fable 5.1 80.3 Sep 2026 Official leaderboard
16 GLM-5.3 Flash-X 80.1 Official leaderboard
17 GPT-5.6 Sol 77.6 Jul 2026 Official leaderboard
18 Muse Wonder 76.2 Oct 2026 Official leaderboard
19 Tencent Hy4 preview 74.6 Aug 2026 Official leaderboard
20 Qwen 3.8 Max 72.0 Aug 2026 Official leaderboard
21 Gemini 3.8 Flash 68.9 Sep 2026 Official leaderboard
22 Muse Spark 1.3 66.6 Sep 2026 Official leaderboard
23 Kimi K3 65.7 Jul 2026 Official leaderboard

Model efficiency

Average score against mean completion time per artifact; the page can also plot cache hit rate and token use. The line joins the runs that no other run beats on both measures. We redrew this chart from the numbers on the official leaderboard.

60 70 80 90 100 05001,0001,5002,000 Completion time (s), lower is better Average score GPT-5.6 Sol: Average score 77.6, completion time 193 s Kimi K3: Average score 65.7, completion time 841 s Qwen 3.8 Max: Average score 72, completion time 1,587 s Tencent Hy4 preview: Average score 74.6, completion time 1,752 s Claude Fable 5.1: Average score 80.3, completion time 766 s Gemini 3.8 Flash: Average score 68.9, completion time 227 s Muse Spark 1.3: Average score 66.6, completion time 198 s GPT-6 Astra: Average score 82.7, completion time 663 s DeepSeek V4.1 Flash: Average score 81.2, completion time 316 s Grok 4.7: Average score 82.9, completion time 977 s GPT-6 Luna: Average score 83.1, completion time 255 s GPT-6 Sol: Average score 89.9, completion time 285 s Claude Sonnet 5.5: Average score 86.8, completion time 245 s GPT-6.1 Sol: Average score 90.1, completion time 373 s Gemini 4 Argon: Average score 81.2, completion time 384 s Muse Wonder: Average score 76.2, completion time 199 s MiniMax-M3.1-Flash: Average score 81.7, completion time 1,050 s Tencent Hy4: Average score 90, completion time 912 s MiMo 2.6 Flash: Average score 83.5, completion time 958 s. Not in our model list GLM-5.3 Flash-X: Average score 80.1, completion time 479 s. Not in our model list Claude Haiku 5.5: Average score 82.2, completion time 149 s Kimi K3.1 Pro: Average score 91, completion time 833 s GPT-6.1 Terra: Average score 90.3, completion time 159 s Kimi K3.1 ProGPT-6.1 TerraClaude Haiku 5.5 60 70 80 90 100 05001,0001,5002,000 Completion time (s), lower is better Average score GPT-5.6 Sol: Average score 77.6, completion time 193 s Kimi K3: Average score 65.7, completion time 841 s Qwen 3.8 Max: Average score 72, completion time 1,587 s Tencent Hy4 preview: Average score 74.6, completion time 1,752 s Claude Fable 5.1: Average score 80.3, completion time 766 s Gemini 3.8 Flash: Average score 68.9, completion time 227 s Muse Spark 1.3: Average score 66.6, completion time 198 s GPT-6 Astra: Average score 82.7, completion time 663 s DeepSeek V4.1 Flash: Average score 81.2, completion time 316 s Grok 4.7: Average score 82.9, completion time 977 s GPT-6 Luna: Average score 83.1, completion time 255 s GPT-6 Sol: Average score 89.9, completion time 285 s Claude Sonnet 5.5: Average score 86.8, completion time 245 s GPT-6.1 Sol: Average score 90.1, completion time 373 s Gemini 4 Argon: Average score 81.2, completion time 384 s Muse Wonder: Average score 76.2, completion time 199 s MiniMax-M3.1-Flash: Average score 81.7, completion time 1,050 s Tencent Hy4: Average score 90, completion time 912 s MiMo 2.6 Flash: Average score 83.5, completion time 958 s. Not in our model list GLM-5.3 Flash-X: Average score 80.1, completion time 479 s. Not in our model list Claude Haiku 5.5: Average score 82.2, completion time 149 s Kimi K3.1 Pro: Average score 91, completion time 833 s GPT-6.1 Terra: Average score 90.3, completion time 159 s Kimi K3.1 ProGPT-6.1 TerraClaude Haiku 5.5
ModelSettingAverage scoreCompletion time
Kimi K3.1 Pro 91 832.8 s
GPT-6.1 Terra 90.3 158.7 s
GPT-6.1 Sol 90.1 372.8 s
Tencent Hy4 90 912.1 s
GPT-6 Sol 89.9 284.9 s
Claude Sonnet 5.5 86.8 244.8 s
MiMo 2.6 Flash 83.5 958.1 s
GPT-6 Luna 83.1 254.9 s
Grok 4.7 82.9 977.2 s
GPT-6 Astra 82.7 663 s
Claude Haiku 5.5 82.2 148.8 s
MiniMax-M3.1-Flash 81.7 1,050 s
DeepSeek V4.1 Flash 81.2 316 s
Gemini 4 Argon 81.2 383.7 s
Claude Fable 5.1 80.3 766 s
GLM-5.3 Flash-X 80.1 478.6 s
GPT-5.6 Sol 77.6 193 s
Muse Wonder 76.2 198.6 s
Tencent Hy4 preview 74.6 1,752 s
Qwen 3.8 Max 72 1,587 s
Gemini 3.8 Flash 68.9 227 s
Muse Spark 1.3 66.6 198 s
Kimi K3 65.7 841 s

The best score, 91.0, costs $1.46 a task

Runs on the cost frontier score higher than any run that's cheaper. GPT-6.1 Terra comes within 0.7 of the best for $0.12 a task, 12 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

OpenAI Moonshot AI Cost frontier One model at each effort level
0.00 20.0 40.0 60.0 80.0 100.0 $0.01 $0.10 $1.00 $10 Cost per task, log scale Kimi K3.1 Pro GPT-6.1 Terra GPT-6 Luna 0.00 20.0 40.0 60.0 80.0 100.0 $0.01 $0.10 $1.00 $10 Cost per task, log scale Kimi K3.1 Pro GPT-6.1 Terra

Source: Official OpenDesign Arena leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Kimi K3.1 Pro 91.0 $1.46 32,679
GPT-6.1 Terra 90.3 $0.12 11,426
GPT-6.1 Sol 90.1 $0.38 23,311
Tencent Hy4 90.0 $0.23 101,426
GPT-6 Sol 89.9 $0.30 13,706
Claude Sonnet 5.5 86.8 $0.57 31,318
GPT-6 Luna 83.1 $0.02 17,265
Grok 4.7 82.9 $0.66 44,268
GPT-6 Astra 82.7 $1.61 15,544
Claude Haiku 5.5 82.2 $0.02 24,760
MiniMax-M3.1-Flash 81.7 $0.25 126,582
DeepSeek V4.1 Flash 81.2 $0.02 29,202
Gemini 4 Argon 81.2 $0.67 41,624
Claude Fable 5.1 80.3 $3.66 63,610
GPT-5.6 Sol 77.6 $0.54 14,521
Muse Wonder 76.2 $1.25 21,235
Tencent Hy4 preview 74.6 $0.42 45,980
Qwen 3.8 Max 72.0 $0.60 28,629
Gemini 3.8 Flash 68.9 $0.21 27,851
Muse Spark 1.3 66.6 $0.27 13,428
Kimi K3 65.7 $0.61 16,066

7 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI Anthropic xAI Moonshot AI
Own leaderboard 7 leaders
Kimi K3.1 Pro 91.0

Source: Official OpenDesign Arena leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
60.0 70.0 80.0 90.0 100.0 2027 60.0 70.0 80.0 90.0 100.0 2027

Source: Official OpenDesign Arena leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Coding

All 282 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471