FrontierSWE
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.
| # | Modell | Organisation | Ergebnis | Nachweis |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 (max) API access | Anthropic | 69,9 % | – |
| 2 | GPT-6 Astra (max) API access | OpenAI | 65,5 % | – |
| 3 | Claude Opus 5.5 (max) API access | Anthropic | 62,3 % | – |
| 4 | Claude Sonnet 5.5 (max) API access | Anthropic | 61,9 % | – |
| 5 | Gemini 4 Carbon (high) API access | Google DeepMind | 61,2 % | – |
| 6 | Claude Fable 5.1 (max) API access | Anthropic | 56,3 % | – |
| 7 | Gemini 4 Argon (high) API access | Google DeepMind | 55 % | – |
| 8 | Claude Opus 5 (max) API access | Anthropic | 52 % | – |
| 9 | Claude Fable 5 (max) API access | Anthropic | 47 % | – |
| 10 | Claude Haiku 5.5 (max) API access | Anthropic | 43,8 % | – |
| 11 | GPT-5.6 Sol (max) API access | OpenAI | 32,2 % | – |
| 12 | GLM-5.3 (max) Open weights (restricted use) | Z.ai (Zhipu AI) | 30,2 % | – |
| 13 | Grok 4.7 (xhigh) API access | xAI | 29,5 % | – |
| 14 | Kimi K3 (max) Open weights (non-commercial) | Moonshot | 25,9 % | – |
| 15 | Muse Spark 1.3 (max) API access | Meta AI | 25,8 % | – |
| 16 | Grok 4.6 (xhigh) API access | xAI | 25,3 % | – |
| 17 | Gemini 3.7 Flash (high) API access | Google DeepMind | 20,3 % | – |
| 18 | Gemini 3.8 Flash (high) API access | Google DeepMind | 19,6 % | – |
| 19 | GLM-5.3-Flash (max) Open weights (unrestricted) | Z.ai (Zhipu AI) | 18,1 % | – |
| 20 | Qwen3.8 Max (0902) (xhigh) API access | Alibaba | 17,8 % | – |
| 21 | Qwen3.8 Max (xhigh) API access | Alibaba | 15,8 % | – |
| 22 | Muse Spark 1.2 (xhigh) API access | Meta AI | 12 % | – |
| 23 | Inkling (xhigh) Open weights (unrestricted) | Thinking Machines | 4,1 % | – |
