Surface Evolver Bench
Das Ergebnis jedes Modells, so, wie der Benchmarking Hub von Epoch AI es veröffentlicht, von Epoch gemessen oder aus der genannten Quelle gesammelt. Facthory hat sie nicht selbst gemessen.
Daten: Epoch AI, „Capabilities & benchmarking“, epoch.ai/benchmarks, lizenziert unter CC BY 4.0. Gelesen am 16.11.2026.
Das Diagramm zeigt die 25 besten von 29 Ergebnissen; die Tabelle führt alle auf.
| # | Modell | Organisation | Ergebnis | Nachweis |
|---|---|---|---|---|
| 1 | Claude Fable 5 (high) API access | Anthropic | 95 % | – |
| 2 | Kimi K3 (unknown) Open weights (non-commercial) | Moonshot | 95 % | – |
| 3 | GPT-5.6 Sol (xhigh) API access | OpenAI | 93,1 % | – |
| 4 | Kimi K3 (max) Open weights (non-commercial) | Moonshot | 93 % | – |
| 5 | GPT-5.5 (high) API access | OpenAI | 88,1 % | – |
| 6 | Claude Opus 4.8 API access | Anthropic | 87,5 % | – |
| 7 | GPT-5.6 Terra (xhigh) API access | OpenAI | 83,8 % | – |
| 8 | GPT-5.5 (medium) API access | OpenAI | 81,3 % | – |
| 9 | Gemini 3.8 Flash (high) API access | Google DeepMind | 76,9 % | – |
| 10 | Grok 4.5 (high) API access | xAI | 74,4 % | – |
| 11 | Claude Opus 4.8 API access | Anthropic | 68,1 % | – |
| 12 | GPT-5.6 Luna (medium) API access | OpenAI | 61,9 % | – |
| 13 | Claude Sonnet 5 (medium) API access | Anthropic | 60 % | – |
| 14 | Gemini 3.5 Flash API access | Google DeepMind | 58,1 % | – |
| 15 | GLM-5.2 Open weights (unrestricted) | Z.ai (Zhipu AI) | 55,6 % | – |
| 16 | MiniMax-M3 Open weights (restricted use) | MiniMax | 55 % | – |
| 17 | GLM-5.3-Flash (max) Open weights (unrestricted) | Z.ai (Zhipu AI) | 52,5 % | – |
| 18 | Muse Spark 1.1 (high) API access | Meta AI | 52,5 % | – |
| 19 | Kimi K2.7 Code Open weights (unrestricted) | Moonshot | 48,8 % | – |
| 20 | DeepSeek V4.1 Flash (high) Open weights (unrestricted) | DeepSeek | 46,3 % | – |
| 21 | Qwen3.8 27B (xhigh) Open weights (unrestricted) | Alibaba | 45 % | – |
| 22 | Qwen3.6 35B-A3B Open weights (unrestricted) | Alibaba | 44,4 % | – |
| 23 | DeepSeek v4 Pro (high) Open weights (unrestricted) | DeepSeek | 40 % | – |
| 24 | Gemma 4 31B IT Open weights (restricted use) | Google DeepMind | 30,6 % | – |
| 25 | Mistral Large 4 Preview (unknown) Open weights (restricted use) | Mistral AI | 28,1 % | – |
| 26 | Mistral Medium 3.5 Open weights (restricted use) | Mistral AI | 26,9 % | – |
| 27 | gpt-oss-120b Open weights (unrestricted) | OpenAI | 25 % | – |
| 28 | Laguna M.1 | Poolside | 15,6 % | – |
| 29 | Trinity Large Thinking Open weights (unrestricted) | Arcee AI | 15,6 % | – |
