| 1 |  Claude Fable 5 | Anthropic | 71.8 | 0 | XHighMediumNone | XHigh63%/37%/0%n 8Medium88%/0%/12%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No None setting | DetailsCharts |
| 2 |  GPT-5.4 | OpenAI | 55.1 | 0 | XHighMediumNone | XHigh88%/12%/0%n 8Medium88%/12%/0%n 8None75%/25%/0%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverage | DetailsCharts |
| 3 | GPT-5.6 Sol | OpenAI | 54.0 | 0 | XHighMediumNone | XHigh100%/0%/0%n 8Medium78%/11%/11%n 9None89%/0%/11%n 9 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-13 | Complete coverage | DetailsCharts |
| 4 | Claude Opus 4.8 | Anthropic | 51.1 | 0 | XHighMediumNone | XHigh75%/13%/12%n 8Medium100%/0%/0%n 8None88%/0%/12%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverage | DetailsCharts |
| 5 | O3 | OpenAI | 49.8 | 0 | XHighMediumNone | XHigh50%/0%/50%n 8Medium75%/0%/25%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No None settingHigh failure rate | DetailsCharts |
| 6 | GPT-5.5 | OpenAI | 48.9 | 0 | XHighMediumNone | XHigh88%/0%/12%n 8Medium88%/12%/0%n 8None88%/0%/12%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverage | DetailsCharts |
| 7 |  Gemini 3.5 Flash | Google | 44.4 | 0 | XHighMediumNone | XHigh0%/75%/25%n 8Medium0%/100%/0%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No None settingHigh repair rateHigh failure rate | DetailsCharts |
| 8 | GPT 5 | OpenAI | 42.0 | 0 | XHighMediumNone | XHigh75%/25%/0%n 8Medium75%/13%/12%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No None setting | DetailsCharts |
| 9 | Claude Opus 4.5 | Anthropic | 41.9 | 0 | XHighMediumNone | XHigh75%/13%/12%n 8Medium100%/0%/0%n 8None100%/0%/0%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverage | DetailsCharts |
| 10 | GPT-5.6 Terra | OpenAI | 40.9 | 0 | XHighMediumNone | XHigh63%/37%/0%n 8Medium63%/37%/0%n 8None78%/11%/11%n 9 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-13 | Complete coverage | DetailsCharts |
| 11 |  DeepSeek V4 Pro | DeepSeek | 39.8 | 0 | XHighMediumNone | XHigh13%/62%/25%n 8None50%/38%/12%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-04 | No Medium settingHigh repair rateHigh failure rate | DetailsCharts |
| 12 | Claude Sonnet 5 | Anthropic | 39.6 | 0 | XHighMediumNone | XHigh100%/0%/0%n 8Medium75%/13%/12%n 8None100%/0%/0%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverage | DetailsCharts |
| 13 |  Kimi K2.7 Code | Moonshot | 39.2 | 0 | XHighMediumNone | Medium0%/75%/25%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingNo None settingHigh repair rateHigh failure rate | DetailsCharts |
| 14 |  MiMo-V2.5-Pro | Xiaomi | 36.8 | 0 | XHighMediumNone | Medium0%/75%/25%n 8None57%/29%/14%n 7 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingHigh repair rateHigh failure rate | DetailsCharts |
| 15 |  GLM-5.2 | Zhipu | 36.6 | 0 | XHighMediumNone | XHigh4%/68%/28%n 25Medium13%/54%/33%n 24None50%/38%/12%n 32 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverageHigh repair rateHigh failure rate | DetailsCharts |
| 16 | GPT 4.1 | OpenAI | 36.6 | 0 | XHighMediumNone | None63%/37%/0%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-04 | No XHigh settingNo Medium setting | DetailsCharts |
| 17 | GPT-5.4 Mini | OpenAI | 35.3 | 0 | XHighMediumNone | XHigh83%/17%/0%n 12Medium92%/4%/4%n 24None79%/21%/0%n 24 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverage | DetailsCharts |
| 18 |  Qwen3.7 Max | Alibaba | 34.9 | 0 | XHighMediumNone | Medium13%/75%/12%n 8None88%/0%/12%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingHigh repair rate | DetailsCharts |
| 19 | GPT-5.4 Nano | OpenAI | 34.0 | 0 | XHighMediumNone | XHigh59%/29%/12%n 17Medium41%/37%/22%n 32None31%/47%/22%n 32 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverage | DetailsCharts |
| 20 | GPT-5.6 Luna | OpenAI | 32.0 | 0 | XHighMediumNone | XHigh78%/11%/11%n 9Medium89%/0%/11%n 9None88%/12%/0%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-13 | Complete coverage | DetailsCharts |
| 21 |  Grok Build 0.1 | xAI | 31.7 | 0 | XHighMediumNone | Medium0%/100%/0%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingNo None settingHigh repair rate | DetailsCharts |
| 22 |  Minimax M3 | MiniMax | 29.2 | 0 | XHighMediumNone | Medium0%/88%/12%n 8None50%/25%/25%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingHigh repair rateHigh failure rate | DetailsCharts |
| 23 | Qwen3.7 Plus | Alibaba | 29.0 | 0 | XHighMediumNone | Medium13%/62%/25%n 8None45%/22%/33%n 9 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingHigh repair rateHigh failure rate | DetailsCharts |
| 24 |  Hy3 Preview | Tencent | 28.5 | 0 | XHighMediumNone | XHigh56%/22%/22%n 9None50%/25%/25%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-04 | No Medium settingHigh failure rate | DetailsCharts |
| 25 | MiMo-V2.5 | Xiaomi | 27.7 | 0 | XHighMediumNone | XHigh0%/63%/37%n 8Medium5%/67%/28%n 21None50%/21%/29%n 24 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverageHigh repair rateHigh failure rate | DetailsCharts |
| 26 |  Nemotron 3 Ultra 550B A55B | NVIDIA | 25.8 | 0 | XHighMediumNone | XHigh12%/44%/44%n 16Medium0%/56%/44%n 9None56%/22%/22%n 9 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverageHigh repair rateHigh failure rate | DetailsCharts |
| 27 | Gemma 4 31B | Google | 24.6 | 0 | XHighMediumNone | Medium0%/57%/43%n 7None72%/14%/14%n 7 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingHigh repair rateHigh failure rate | DetailsCharts |
| 28 |  Step 3.7 Flash | StepFun | 24.1 | 0 | XHighMediumNone | XHigh0%/63%/37%n 8Medium0%/50%/50%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No None settingHigh repair rateHigh failure rate | DetailsCharts |
| 29 | DeepSeek V4 Flash | DeepSeek | 23.5 | 0 | XHighMediumNone | XHigh33%/40%/27%n 15Medium63%/25%/12%n 16None58%/21%/21%n 24 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverageHigh failure rate | DetailsCharts |
| 30 | GPT-OSS 120B | OpenAI | 23.5 | 0 | XHighMediumNone | XHigh0%/63%/37%n 8Medium13%/75%/12%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No None settingHigh repair rateHigh failure rate | DetailsCharts |
| 31 | Gemini 3.1 Flash Lite | Google | 17.0 | 0 | XHighMediumNone | XHigh0%/63%/37%n 8Medium25%/50%/25%n 8None38%/25%/37%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | Complete coverageHigh repair rateHigh failure rate | DetailsCharts |
| 32 |  Nex N2 Pro | Nex AGI | 26.1 | Baseline | XHighMediumNone | Medium0%/63%/37%n 8None0%/44%/56%n 16 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingHigh repair rateHigh failure rate | DetailsCharts |
| 33 |  North Mini Code | Other | 22.7 | Baseline | XHighMediumNone | Medium38%/37%/25%n 8None19%/12%/69%n 16 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-05 | No XHigh settingHigh failure rate | DetailsCharts |
| 34 |  Mistral Medium 3.5 | Mistral | 15.3 | Baseline | XHighMediumNone | XHigh22%/22%/56%n 9None25%/25%/50%n 8 | Overall Models: combines public reasoning variantsStandard GameBench code-generation prompt | 2026-07-04 | No Medium settingHigh failure rate | DetailsCharts |