Benchmark Almanac

DesignArena Agentic Game Dev

Gemini 4 Argon 1,265 leads since September 2026
Not in model cards

Coding agents build multi-file games from a user prompt. Users play the final games and vote, and the logs keep tool calls, retries and failures.

One of 40 DesignArena boards. DesignArena ranks entries with Bradley-Terry ratings from user votes.

Kind
Building whole projects
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Strongest model tested
Gemini 4 Argon #9 on the capability index
Last model added
September 2026
Scores refreshed
November 12th, 2026
27
Models scored on its own leaderboard

Gemini 4 Argon leads with 1,265

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.

27 models. Scroll sideways for the rest. Bars start at 900.

  1. 1,265 Gemini 4 Argon
  2. 1,233 Claude Fable 5
  3. 1,213 GPT-5.6 Sol
  4. 1,204 Claude Opus 5
  5. 1,203 Kimi K3
  6. 1,184 Claude Opus 4.8
  7. 1,184 Grok 4.5
  8. 1,182 Gemini 3.7 Flash
  9. 1,171 Claude Sonnet 5
  10. 1,158 GLM-5.2
  11. 1,158 Grok 4.6
  12. 1,146 Claude Opus 4.6
  13. 1,132 Muse Spark 1.1
  14. 1,129 Gemini 3.5 Flash
  15. 1,127 GPT-5.5
  16. 1,127 Muse Spark 1.2
  17. 1,126 Gemini 3.6 Flash
  18. 1,119 Claude Sonnet 4.6
  19. 1,115 MiniMax-M3
  20. 1,111 GLM-5.1
  21. 1,106 Qwen3.7-Max
  22. 1,089 Kimi K2.6
  23. 1,078 Kimi K2.7 Code
  24. 1,062 Gemini 3.1 Pro
  25. 1,050 GLM-5V-Turbo
  26. 1,032 GPT-5.6 Terra
  27. 956 Grok 4.3

Bars start at 900.

  1. 1 Gemini 4 Argon 1,265
  2. 2 Claude Fable 5 1,233
  3. 3 GPT-5.6 Sol medium, xhigh 1,213
  4. 4 Claude Opus 5 1,204
  5. 5 Kimi K3 1,203
  6. 6 Claude Opus 4.8 1,184
  7. 7 Grok 4.5 1,184
  8. 8 Gemini 3.7 Flash 1,182
  9. 9 Claude Sonnet 5 1,171
  10. 10 GLM-5.2 1,158
  11. 11 Grok 4.6 1,158
  12. 12 Claude Opus 4.6 1,146
  13. 13 Muse Spark 1.1 1,132
  14. 14 Gemini 3.5 Flash 1,129
  15. 15 GPT-5.5 1,127
  16. 16 Muse Spark 1.2 1,127
  17. 17 Gemini 3.6 Flash 1,126
  18. 18 Claude Sonnet 4.6 1,119
  19. 19 MiniMax-M3 1,115
  20. 20 GLM-5.1 1,111
  21. 21 Qwen3.7-Max 1,106
  22. 22 Kimi K2.6 1,089
  23. 23 Kimi K2.7 Code 1,078
  24. 24 Gemini 3.1 Pro 1,062
  25. 25 GLM-5V-Turbo 1,050
  26. 26 GPT-5.6 Terra 1,032
  27. 27 Grok 4.3 956

Source: Official DesignArena Agentic Game Dev leaderboard

#ModelScoreSettingDateSource
1 Gemini 4 Argon 1,265 Sep 2026 Official leaderboard
2 Claude Fable 5 1,233 Jun 2026 Official leaderboard
3 GPT-5.6 Sol 1,213 medium, xhigh Jul 2026 Official leaderboard
4 Claude Opus 5 1,204 Jul 2026 Official leaderboard
5 Kimi K3 1,203 Jul 2026 Official leaderboard
6 Claude Opus 4.8 1,184 May 2026 Official leaderboard
7 Grok 4.5 1,184 Jul 2026 Official leaderboard
8 Gemini 3.7 Flash 1,182 Aug 2026 Official leaderboard
9 Claude Sonnet 5 1,171 Jun 2026 Official leaderboard
10 GLM-5.2 1,158 Jun 2026 Official leaderboard
11 Grok 4.6 1,158 Aug 2026 Official leaderboard
12 Claude Opus 4.6 1,146 Feb 2026 Official leaderboard
13 Muse Spark 1.1 1,132 Jul 2026 Official leaderboard
14 Gemini 3.5 Flash 1,129 May 2026 Official leaderboard
15 GPT-5.5 1,127 Apr 2026 Official leaderboard
16 Muse Spark 1.2 1,127 Aug 2026 Official leaderboard
17 Gemini 3.6 Flash 1,126 Jul 2026 Official leaderboard
18 Claude Sonnet 4.6 1,119 Feb 2026 Official leaderboard
19 MiniMax-M3 1,115 Jun 2026 Official leaderboard
20 GLM-5.1 1,111 Apr 2026 Official leaderboard
21 Qwen3.7-Max 1,106 May 2026 Official leaderboard
22 Kimi K2.6 1,089 Apr 2026 Official leaderboard
23 Kimi K2.7 Code 1,078 Jun 2026 Official leaderboard
24 Gemini 3.1 Pro 1,062 Feb 2026 Official leaderboard
25 GLM-5V-Turbo 1,050 Apr 2026 Official leaderboard
26 GPT-5.6 Terra 1,032 Jul 2026 Official leaderboard
27 Grok 4.3 956 Apr 2026 Official leaderboard

4 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.

Anthropic Google DeepMind
Own leaderboard 4 leaders
Gemini 4 Argon 1,265

Source: Official DesignArena Agentic Game Dev 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
900 1,000 1,100 1,200 1,300 2026 900 1,000 1,100 1,200 1,300 2026

Source: Official DesignArena Agentic Game Dev 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