A contamination-resistant benchmark designed to provide frequently updated and objective evaluation of large language models.
Browse the latest scores, model modes, release dates, and parameter sizes for LiveBench.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
| Rank | Model | License | |||
|---|---|---|---|---|---|
 GPT-5.5 Deep Thinking Mode | 80.71 | 2026-04-23 | Unknown | Closed | |
GPT-5.4 Deep Thinking Mode | 80.28 | 2026-03-05 | Unknown | Closed | |
 Gemini 3.1 Pro Preview Thinking Level · High | 79.93 | 2026-02-20 | Unknown | Closed | |
4 |  Claude Opus 4.8 Deep Thinking Mode | 78.79 | 2026-05-28 | Unknown | Closed |
5 | Claude Fable 5 Deep Thinking Mode | 78.31 | 2026-06-09 | Unknown | Closed |
6 | Claude Opus 4.8 Thinking Level · High | 77.16 | 2026-05-28 | Unknown | Closed |
7 | Opus 4.7 Deep Thinking Mode | 76.91 | 2026-04-16 | Unknown | Closed |
8 | Claude Opus 4.6 Thinking Level · High | 76.33 | 2026-02-05 | Unknown | Closed |
9 | GPT-5.5 Thinking Level · High | 76.24 | 2026-04-23 | Unknown | Closed |
10 |  GLM-5.2 Standard Mode | 76.24 | 2026-06-13 | 753.3B | Free Commercial |
11 | Opus 4.5 64K | 75.96 | 2025-11-25 | Unknown | Closed |
12 | Claude Sonnet 4.6 Thinking Level · Medium | 75.47 | 2026-02-17 | Unknown | Closed |
13 | Claude Opus 4.8 Thinking Level · Medium | 75.47 | 2026-05-28 | Unknown | Closed |
14 | Claude Fable 5 Thinking Level · High | 75.47 | 2026-06-09 | Unknown | Closed |
15 | Claude Sonnet 4.6 Thinking Level · High | 75.32 | 2026-02-17 | Unknown | Closed |
16 | GPT-5.4 Thinking Level · High | 75.07 | 2026-03-05 | Unknown | Closed |
17 | Gemini 3.5 Flash Thinking Level · High | 75.02 | 2026-06-20 | Unknown | Closed |
18 | Opus 4.7 Thinking Level · High | 74.89 | 2026-04-16 | Unknown | Closed |
19 | GPT-5.2 Thinking Level · High | 74.84 | 2025-12-11 | Unknown | Closed |
20 | GPT-5.2-Codex Standard Mode | 74.30 | 2025-12-18 | Unknown | Closed |
21 |  Qwen3.7 Max Deep Thinking Mode | 74.29 | 2026-05-20 | Unknown | Closed |
22 | GPT-5.1-Codex-Max Deep Thinking Mode | 73.98 | 2025-11-19 | Unknown | Closed |
23 |  DeepSeek-V4-Pro Standard Mode | 73.58 | 2026-04-24 | 1600B | Free Commercial |
24 | Gemini 3.0 Pro (Preview 11-2025) Thinking Level · High | 73.39 | 2025-11-18 | Unknown | Closed |
25 | GPT-5.3 Codex Thinking Level · High | 72.76 | 2026-02-05 | Unknown | Closed |
26 | Gemini 3.0 Flash Thinking Level · High | 72.40 | 2025-12-17 | Unknown | Closed |
27 | Opus 4.7 Thinking Level · Medium | 72.31 | 2026-04-16 | Unknown | Closed |
28 |  Kimi K2.6 Thinking Enabled | 72.17 | 2026-04-20 | 1000B | Free Commercial |
29 | GPT-5.1 Thinking Level · High | 72.04 | 2025-11-12 | Unknown | Closed |
30 | Kimi K2.7 Code Standard Mode | 71.89 | 2026-06-12 | 1000B | Free Commercial |
