| Modellslot | Typ | Version | AA Index | Arena Elo | SWE-bench | Aider | LiveBench | BFCL | RULER | MMMU | GPQA | Bemerkung |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Claude Opus Anthropic | Cloud | Opus 4.6 | 100% | 98% | 98% | 96% | 99% | 96% | 94% | 91% | 99% | Beispieldaten |
GPT Flagship OpenAI | Cloud | GPT-5.5 | 98% | 100% | 96% | 94% | 100% | 97% | 91% | 95% | 100% | Beispieldaten |
Gemini Pro Google | Cloud | Gemini 3 Pro | 94% | 96% | 89% | 86% | 96% | 89% | 100% | 98% | 96% | Beispieldaten |
Claude Sonnet Anthropic | Cloud | Sonnet 4.6 | 91% | 94% | 95% | 93% | 88% | 90% | 86% | 84% | 88% | Beispieldaten |
Qwen Coder Alibaba / Open Weights | Open | Qwen Coder 3.x | 82% | 84% | 89% | 90% | 73% | 70% | 78% | – | 73% | Beispieldaten |
DeepSeek Reasoner DeepSeek / Open Weights | Open/API | DeepSeek R2 | 80% | 82% | 76% | 72% | 88% | 66% | 74% | – | 88% | Beispieldaten |
Llama Instruct Meta / Open Weights | Open | Llama 4.x | 74% | 78% | 70% | 68% | 76% | 64% | 82% | 62% | 76% | Beispieldaten |
Mistral Large Mistral | Hybrid | Large 3 | 72% | 76% | 68% | 66% | 73% | 70% | 76% | 58% | 73% | Beispieldaten |
Gemma Google / Open Weights | Open | Gemma 4 12B | 66% | 70% | 55% | 51% | 68% | 50% | 72% | 78% | 68% | Beispieldaten |
Qwen Consumer Laptop/MacBook class | Local | Qwen 3 8B GGUF | 58% | 63% | 52% | 49% | 61% | 44% | 67% | – | 61% | Beispieldaten |