PurrBench: Measuring AI Purrformance
Task: Cat RPG · models: 87 · runs: 121 · head-to-head comparisons: 0 · games submitted: 74.
Ranking
Time = architecture + game; architecture only if no game was submitted. Time does not affect the overall score.
| # | Model▼ run conditions | OVERALL▼ | Architecture▼ | Gamedev▼ own · given | Frontend▼ own · given | Other tasks▼ | Time▼ | KB▼ arch. / own / given | Play |
|---|---|---|---|---|---|---|---|---|---|
| 1 | GPT-6 AstraproChatGPT | 77.577.5 | 66.8 | 88.4/ | 87.6/ | 32:43 | 15/110/— | ▶ | |
| 2 | GPT-6 AstralowAPI codex | 72.672.6 | 67.2 | 78.2/ | 77.3/ | 10:13 | 35/71/— | ▶ | |
| 3 | GPT-6 AstramediumAPI codex | 72.672.6 | 66.9 | 77.5/ | 81.4/ | 18:31 | 35/80/— | ▶ | |
| 4 | Sakana Fugu Ultra v2maxAPI openrouter | 71.171.1 | 71.1 | / | / | 15:52 | 69/—/— | ||
| 5 | GPT-5.6 SolproChatGPT | 71.070.5 | 67.0 | 74.0/71.6 | 78.7/78.1 | 152:20 | 71/109/104 | ▶▶ | |
| 6 | Grok 4.7highAPI openrouter | 69.669.6 | 66.4 | 72.0/ | 76.4/ | 29:45 | 34/119/— | ▶ | |
| 7 | GPT-6 SolmediumAPI codex | 68.768.7 | 56.5 | 80.8/ | 81.5/ | 3:55 | 9.2/27/— | ▶ | |
| 8 | Claude Opus 5highclaude code | 68.768.1 | 60.2 | 77.4/75.1 | 76.3/73.6 | 77:43 | 28/195/437 | ▶▶ | |
| 9 | GPT-6 Astrahighcodex | 68.168.1 | 54.3 | 81.4/ | 83.8/ | 24:55 | 7.4/85/— | ▶ | |
| 10 | Claude Opus 5xhighclaude code | 67.267.2 | 67.2 | / | / | 27:15 | 177/—/— | ||
| 11 | GPT-6 Solmediumcodex | 66.966.9 | 51.7 | 81.5/ | 84.5/ | 11:20 | 5.6/33/— | ▶ | |
| 12 | GPT-6 AstraxhighAPI codex | 66.566.5 | 66.5 | / | / | 5:08 | 38/—/— | ||
| 13 | GPT-6 Astramediumcodex | 68.866.3 | 53.7 | 83.9/73.1 | 83.8/78.1 | 13:27 | 6.2/50/134 | ▶▶ | |
| 14 | Kimi K3maxAPI opencode-go | 66.266.2 | 60.9 | 71.5/ | 71.8/ | 24:15 | 29/96/— | ▶ | |
| 15 | HY4 Previewmaxopencode | 66.166.1 | 59.0 | 72.6/ | 75.2/ | 70:26 | 21/183/— | ▶ | |
| 16 | GPT-6 Astralowcodex | 68.066.0 | 52.3 | 83.5/74.6 | 84.2/80.9 | 13:19 | 4.9/46/120 | ▶▶ | |
| 17 | Claude Fable 5.1mediumclaude code | 66.066.0 | 57.1 | 76.3/ | 69.3/ | 56:06 | 16/96/— | ▶ | |
| 18 | Union AlphaAPI openrouter | 66.066.0 | 66.0 | / | / | 3:20 | 27/—/— | ||
| 19 | Claude Fable 5maxClaude.ai | 65.465.4 | 65.4 | / | / | 8:41 | 38/—/— | ||
| 20 | GPT-5.6 Solmediumcodex | 69.164.2 | 58.1 | 79.1/58.7 | 83.9/68.8 | 11:08 | 12/31/51 | ▶▶ | |
| 21 | Claude Opus 5mediumclaude code | 65.064.0 | 55.2 | 75.0/69.9 | 74.1/73.4 | 32:34 | 17/140/427 | ▶▶ | |
| 22 | Sakana FugumaxAPI openrouter | 63.763.7 | 54.5 | 73.2/ | 72.0/ | 23:29 | 45/96/— | ▶ | |
| 23 | Claude Opus 5.5highclaude code | 63.663.6 | 50.2 | 78.3/ | 71.6/ | 67:07 | 15/248/— | ▶ | |
| 24 | Grok 4.6highCursor | 65.463.4 | 60.3 | 69.6/61.8 | 73.7/65.9 | 19:37 | 19/68/111 | ▶▶ | |
| 25 | GLM 5.3low variantopencode | 63.263.2 | 55.7 | 70.5/ | 71.6/ | 10:42 | 9.9/63/— | ▶ | |
| 26 | Claude Sonnet 5highclaude code | 67.963.0 | 56.6 | 79.6/58.5 | 77.2/64.0 | 38:12 | 20/73/240 | ▶▶ | |
| 27 | GPT-5.6 SolmediumDevin | 67.863.0 | 59.4 | 75.8/55.2 | 78.3/62.8 | 5:55 | 13/38/39 | ▶▶ | |
| 28 | GPT-6 AstrahighAPI codex | 63.063.0 | 63.0 | / | / | 3:07 | 27/—/— | ||
| 29 | Kimi K3opencode | 63.262.0 | 55.5 | 70.4/65.2 | 73.0/68.9 | 10:12 | 10/48/289 | ▶▶ | |
| 30 | Claude Opus 5.5mediumclaude code | 61.961.9 | 48.9 | 75.7/ | 71.9/ | 49:01 | 12/115/— | ▶ | |
| 31 | GPT-5.6 SolmaxAPI codex | 61.961.9 | 61.9 | / | / | 5:29 | 41/—/— | ||
| 32 | GPT-5.6 SolmediumAPI codex | 65.161.8 | 58.5 | 70.4/56.7 | 76.6/66.1 | 8:45 | 42/55/61 | ▶▶ | |
| 33 | Muse Spark 1.3xhighopencode | 61.361.3 | 51.8 | 70.6/ | 71.6/ | 4:11 | 7.9/48/— | ▶ | |
| 34 | DeepSeek V4.1 FlashmaxAPI openrouter | 61.061.0 | 61.4 | 60.0/ | 63.1/ | 7:10 | 44/102/— | ▶ | |
| 35 | Grok 4.5highCursor | 60.760.7 | 50.9 | 70.1/ | 72.5/ | 3:11 | 10.0/53/— | ▶ | |
| 36 | Grok 4.6highAPI openrouter | 60.760.7 | 56.9 | 63.7/ | 68.1/ | 19:32 | 39/89/— | ▶ | |
| 37 | GPT-6 LunamediumAPI codex | 60.160.1 | 54.4 | 65.6/ | 67.0/ | 2:40 | 13/19/— | ▶ | |
| 38 | GPT-5.5highAPI codex | 59.659.6 | 52.7 | 66.2/ | 68.2/ | 7:56 | 25/46/— | ▶ | |
| 39 | DeepSeek V4.1 FlashmediumAPI openrouter | 58.558.5 | 51.4 | 64.9/ | 68.1/ | 2:25 | 17/64/— | ▶ | |
| 40 | Qwen 3.8maxopencode | 58.558.5 | 53.9 | 63.1/ | 62.8/ | 46:20 | 15/78/— | ▶ | |
| 41 | GPT-5.5xhighcodex | 62.058.3 | 48.8 | 74.5/59.3 | 77.7/64.2 | 14:30 | 8.0/61/110 | ▶▶ | |
| 42 | Muse Spark 1.2opencode | 58.358.3 | 53.3 | 62.6/ | 66.1/ | 2:37 | 11/50/— | ▶ | |
| 43 | SWE-2highDevin | 60.158.2 | 50.4 | 69.7/62.7 | 70.1/59.4 | 53:04 | 6.5/49/265 | ▶▶ | |
| 44 | Kimi K3highAPI synthetic | 57.957.9 | 57.6 | 58.1/ | 58.5/ | 10:58 | 19/69/— | ▶ | |
| 45 | SWE-2maxDevin | 63.257.8 | 50.2 | 75.8/53.7 | 77.9/58.5 | 67:03 | 6.9/64/354 | ▶▶ | |
| 46 | GPT-5.5mediumcodex | 62.157.6 | 53.0 | 70.3/51.6 | 75.1/60.0 | 6:47 | 7.7/35/32 | ▶▶ | |
| 47 | GPT-6 Lunamediumcodex | 56.556.5 | 47.4 | 64.3/ | 70.7/ | 3:04 | 5.2/16/— | ▶ | |
| 48 | Qwen 3.8 MaxAPI opencode-go | 56.356.3 | 55.5 | 56.7/ | 58.9/ | 16:45 | 18/81/— | ▶ | |
| 49 | Grok 4.6highWeb chat arena.ai | 55.855.8 | 63.1 | 47.8/ | 51.7/ | 12:25 | 19/61/— | ▶ | |
| 50 | Sonnet 5mediumclaude code | 55.955.7 | 50.7 | 61.8/58.8 | 58.7/65.4 | 7:51 | 15/62/121 | ▶▶ | |
| 51 | SWE-2mediumDevin | 55.855.1 | 45.7 | 67.1/62.4 | 61.5/65.4 | 11:55 | 4.1/25/188 | ▶▶ | |
| 52 | Ox AlphamaxAPI openrouter | 54.854.8 | 48.1 | 60.2/ | 66.8/ | 32:36 | 12/85/— | ▶ | |
| 53 | GPT-5.3 SparkhighAPI codex | 54.554.5 | 49.4 | 59.7/ | 59.1/ | 1:23 | 14/56/— | ▶ | |
| 54 | GPT-5.6 SolhighAPI codex | 54.554.5 | 54.5 | / | / | 3:59 | 47/—/— | ||
| 55 | GPT-5.5xhighAPI codex | 54.254.2 | 54.5 | 52.2/ | 60.5/ | 2:37 | 36/65/— | ▶ | |
| 56 | GPT-5.6 SolxhighAPI codex | 54.054.0 | 54.0 | / | / | 3:26 | 41/—/— | ||
| 57 | Ox AlphahighAPI openrouter | 53.953.9 | 45.8 | 62.3/ | 60.7/ | 13:08 | 8.2/47/— | ▶ | |
| 58 | GLM 5.3maxAPI opencode-go | 53.253.2 | 53.2 | / | / | 9:46 | 22/—/— | ||
| 59 | GPT-5.5mediumAPI codex | 58.453.1 | 50.0 | 65.4/43.6 | 72.2/53.0 | 5:38 | 24/47/39 | ▶▶ | |
| 60 | GPT-5.3 Sparkhighcodex | 52.652.6 | 45.4 | 59.6/ | 60.7/ | 0:45 | 5.1/28/— | ▶ | |
| 61 | MiMo V2.6 ProhighAPI openrouter | 52.652.6 | 45.3 | 58.8/ | 64.5/ | 24:17 | 39/77/— | ▶ | |
| 62 | DeepSeek V4 Flashopencode | 52.552.5 | 42.8 | 62.4/ | 61.9/ | 60:50 | 13/68/— | ▶ | |
| 63 | Muse Spark 1.2highAPI opencode-zen | 52.152.1 | 48.0 | 55.2/ | 60.2/ | 5:01 | 10/27/— | ▶ | |
| 64 | Claude Sonnet 5highWeb chat arena.ai | 51.251.2 | 51.2 | / | / | 1:57 | 26/—/— | ||
| 65 | Gemini 3.8 FlashhighCursor | 54.350.9 | 43.4 | 64.0/50.1 | 70.1/57.5 | 6:07 | 22/117/269 | ▶▶ | |
| 66 | Muse Spark 1.2xhighAPI opencode-zen | 50.950.9 | 46.1 | 55.0/ | 58.1/ | 1:21 | 11/28/— | ▶ | |
| 67 | GLM 5.2maxWeb chat arena.ai | 50.750.7 | 43.2 | 57.9/ | 59.0/ | 3:02 | 8.5/42/— | ▶ | |
| 68 | MiMo V2.5highAPI openrouter | 50.750.7 | 45.3 | 54.4/ | 62.8/ | 5:58 | 20/69/— | ▶ | |
| 69 | HY3highAPI opencode-zen | 50.550.5 | 43.8 | 57.9/ | 54.8/ | 3:12 | 5.0/22/— | ▶ | |
| 70 | Space Bunny AlphamaxAPI openrouter | 49.749.7 | 59.1 | 38.5/ | 47.1/ | 7:29 | 43/91/— | ▶ | |
| 71 | Gemini 3.8 FlashmediumCursor | 49.649.6 | 40.3 | 57.8/ | 63.2/ | 3:02 | 19/82/— | ▶ | |
| 72 | Gemini 3.7 FlashhighAPI openrouter | 49.349.3 | 35.9 | 62.1/ | 64.9/ | 2:59 | 14/78/— | ▶ | |
| 73 | PrismML BonsaixhighAPI openrouter | 49.249.2 | 49.2 | / | / | 4:18 | 11/—/— | ||
| 74 | Gemini 3.7 FlashhighCursor | 54.248.5 | 43.4 | 62.8/39.2 | 73.8/53.1 | 4:02 | 18/100/163 | ▶▶ | |
| 75 | Gemini 3.7 FlashhighWeb chat arena.ai | 48.048.0 | 38.4 | 56.4/ | 62.2/ | 1:42 | 13/62/— | ▶ | |
| 76 | DeepSeek V4 ProhighAPI opencode-go | 46.446.4 | 44.3 | 48.1/ | 50.3/ | 5:02 | 16/49/— | ▶ | |
| 77 | Qwen 3.7maxWeb chat arena.ai | 46.146.1 | 44.2 | 48.4/ | 46.7/ | 2:16 | 8.4/32/— | ▶ | |
| 78 | MiMo V2.6 FlashhighAPI openrouter | 45.645.6 | 45.6 | 44.9/ | 48.4/ | 7:03 | 28/59/— | ▶ | |
| 79 | DeepSeek V4 FlashmaxAPI opencode-go | 44.644.6 | 44.3 | 45.2/ | 44.1/ | 1:43 | 11/44/— | ▶ | |
| 80 | Arena Max — model unknownWeb chat arena.ai | 44.144.1 | 44.1 | / | / | 11:45 | 18/—/— | ||
| 81 | Gemma 4 31BhighAPI openrouter | 41.741.7 | 43.3 | 39.6/ | 42.4/ | 1:33 | 4.7/21/— | ▶ | |
| 82 | Kimi K3maxAPI synthetic | 41.641.6 | 54.6 | 25.9/ | 39.8/ | 10:31 | 27/30/— | ▶ | |
| 83 | DeepSeek V4 PromaxAPI opencode-go | 39.039.0 | 40.5 | 35.4/ | 45.4/ | 6:29 | 19/86/— | ▶ | |
| 84 | Claude Haiku 4.5claude code | 38.238.2 | 37.9 | 37.8/ | 41.7/ | 2:14 | 6.0/36/— | ▶ | |
| 85 | Solar Mini 4maxAPI openrouter | 38.238.2 | 42.0 | 32.5/ | 41.6/ | 4:23 | 34/38/— | ▶ | |
| 86 | IBM Granite 4.2 8BmaxAPI openrouter | 35.335.3 | 40.1 | 29.9/ | 32.7/ | 8:35 | 35/13/— | ▶ | |
| 87 | IBM Granite 4.2 8BAPI openrouter | 33.633.6 | 45.8 | 20.7/ | 24.4/ | 1:50 | 22/18/— | ▶ | |
