LiveBench (contamination-free, monthly-updated LLM benchmark) — Aggregate/overall score from Kimi-K2 cards, diff source than livebench_model_judgment entries.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | o3-mini | OpenAI | 84.6 | 1 | 2026-08-23 |
| 2 | GPT-5.5 | OpenAI | 80.7 | 1 | 2026-08-23 |
| 3 | GPT-5.4 | OpenAI | 80.3 | 1 | 2026-08-23 |
| 4 | Gemini 3.1 Pro | 79.9 | 1 | 2026-08-23 | |
| 5 | Claude Fable 5 | Anthropic | 78.3 | 1 | 2026-08-23 |
| 6 | Claude Opus 4.8 | Anthropic | 77.2 | 1 | 2026-08-23 |
| 7 | Qwen3 235B A22B | Alibaba | 77.1 | 3 | 2026-08-23 |
| 8 | Claude Opus 4.7 | Anthropic | 76.9 | 1 | 2026-08-23 |
| 9 | Kimi K2 Instruct | Moonshot | 76.4 | 4 | 2026-08-23 |
| 10 | Claude Opus 4.6 | Anthropic | 76.3 | 1 | 2026-08-23 |
| 11 | Claude Opus 4.5 | Anthropic | 76 | 1 | 2026-08-23 |
| 12 | Claude Sonnet 4.6 | Anthropic | 75.5 | 1 | 2026-08-23 |
| 13 | Gemini 3.5 Flash | 75 | 1 | 2026-08-23 | |
| 14 | Qwen3 32B | Alibaba | 74.9 | 1 | 2026-08-23 |
| 15 | GPT-5.2 | OpenAI | 74.8 | 1 | 2026-08-23 |
| 16 | Claude Sonnet 4 | Anthropic | 74.8 | 2 | 2026-06-15 |
| 17 | Claude 4 Opus | Anthropic | 74.6 | 2 | 2026-06-15 |
| 18 | Qwen3 30B A3B | Alibaba | 74.3 | 1 | 2026-08-23 |
| 19 | GPT-5.2 Codex | OpenAI | 74.3 | 1 | 2026-08-23 |
| 20 | Qwen3.7 Max | Alibaba | 74.3 | 1 | 2026-08-23 |
| 21 | DeepSeek-V4-Pro | DeepSeek | 73.6 | 1 | 2026-08-23 |
| 22 | Gemini 3 Pro | 73.4 | 1 | 2026-08-23 | |
| 23 | QwQ 32B Preview | Alibaba | 73.1 | 1 | 2026-08-23 |
| 24 | GPT-5.3 Codex | OpenAI | 72.8 | 1 | 2026-08-23 |
| 25 | Gemini 3 Flash Preview | 72.4 | 1 | 2026-08-23 | |
| 26 | DeepSeek-V3 | DeepSeek | 72.4 | 2 | 2026-06-15 |
| 27 | Kimi K2.6 | Moonshot | 72.2 | 1 | 2026-08-23 |
| 28 | Kimi K2.7 Code | Moonshot | 71.9 | 1 | 2026-08-23 |
| 29 | Qwen3.6 Plus | Alibaba | 70.8 | 1 | 2026-08-23 |
| 30 | GLM-5.1 | Z.ai | 70.2 | 1 | 2026-08-23 |
| 31 | GPT-5.4 nano | OpenAI | 70.1 | 1 | 2026-08-23 |
| 32 | MiniMax M3 | MiniMax | 70 | 1 | 2026-08-23 |
| 33 | Kimi K2.5 | Moonshot | 69.1 | 1 | 2026-08-23 |
| 34 | Qwen2.5 Instruct 72B | Alibaba | 52.3 | 1 | 2026-08-23 |
| 35 | o1 | OpenAI | 52.3 | 1 | 2026-08-23 |
| 36 | Phi 4 | Microsoft | 47.6 | 1 | 2026-08-23 |
| 37 | Qwen2.5 Omni 7B | Alibaba | 29.6 | 1 | 2026-08-23 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.