🧠 Reasoning · llm-stats.com
GPQA Diamond
Graduate-level science questions in physics, chemistry and biology, designed to be Google-proof.
45
نماذج مختبرة
Claude Mythos Preview
النموذج الأعلى
94.6%
أفضل درجة
Anthropic
المزوّد
| # | النموذج | GPQA Diamond |
|---|---|---|
| 1 | Claude Mythos Preview Anthropic | 94.6% |
| 2 | GPT-5.6 Sol OpenAI | 94.6% |
| 3 | Gemini 3.1 Pro Google | 94.3% |
| 4 | Claude Opus 4.7 Anthropic | 94.2% |
| 5 | Claude Opus 4.8 Anthropic | 93.6% |
| 6 | GPT-5.5 OpenAI | 93.6% |
| 7 | Kimi K3 Moonshot AI | 93.5% |
| 8 | Claude Opus 5 Anthropic | 93.2% |
| 9 | Grok 4.5 xAI | 93% |
| 10 | GPT-5.6 Terra OpenAI | 92.9% |
| 11 | MiniMax M3 MiniMax | 92.9% |
| 12 | Gemini 3.6 Flash Google | 92.8% |
| 13 | GPT-5.4 OpenAI | 92.8% |
| 14 | Claude Fable 5 Anthropic | 92.6% |
| 15 | Qwen3.8 Max Alibaba | 92.6% |
| 16 | Qwen3.7 Max Alibaba | 92.4% |
| 17 | GPT-5.2 OpenAI | 92.4% |
| 18 | GPT-5.6 Luna OpenAI | 92.3% |
| 19 | Gemini 3.5 Flash Google | 92.2% |
| 20 | Gemini 3 Pro Google | 91.9% |
| 21 | GPT-5.3 Codex OpenAI | 91.5% |
| 22 | Claude Opus 4.6 Anthropic | 91.3% |
| 23 | GLM-5.2 Z.ai | 91.2% |
| 24 | Kimi K2.6 Moonshot AI | 90.5% |
| 25 | Gemini 3 Flash Google | 90.4% |
| 26 | Hunyuan Hy3 Tencent | 90.4% |
| 27 | Qwen3.6 Plus Alibaba | 90.4% |
| 28 | Qwen3.7 Plus Alibaba | 90.3% |
| 29 | DeepSeek V4 Pro DeepSeek | 90.1% |
| 30 | Claude Sonnet 4.6 Anthropic | 89.9% |
| 31 | Muse Spark 1.1 Meta | 89.8% |
| 32 | Kimi K2.7 Code Moonshot AI | 89.6% |
| 33 | Muse Spark Meta | 89.5% |
| 34 | Seed 2.0 Pro ByteDance | 88.9% |
| 35 | Grok 4 Heavy xAI | 88.4% |
| 36 | Qwen3.5 397B-A17B Alibaba | 88.4% |
| 37 | GPT-5.1 OpenAI | 88.1% |
| 38 | DeepSeek V4 Flash DeepSeek | 88.1% |
| 39 | Inkling Thinking Machines | 87.2% |
| 40 | Claude Opus 4.5 Anthropic | 87% |
| 41 | Nemotron 3 Ultra 550B NVIDIA | 87% |
| 42 | MiMo V2.5 Pro Xiaomi | 86.6% |
| 43 | Qwen3.5 122B-A10B Alibaba | 86.6% |
| 44 | GPT-5 OpenAI | 85.7% |
| 45 | Claude Sonnet 5 Anthropic | 80% |
بيانات الاختبار مجمّعة من llm-stats.com. اعرف المزيد في المنهجية.