← LiveBench leaderboards

Plot Unscrambling

LiveBench · 28 models

RankModelScore
1claude-fable-5-xhigh-effort76.373
2gpt-5.5-high75.28
3gemini-3.1-pro-preview-high74.128
4gpt-5.5-xhigh74.089
5gpt-5.570.711
6claude-opus-4-8-xhigh-effort70.419
7gemini-3.5-flash-high67.753
8claude-opus-4-6-thinking-auto-high-effort66.476
9claude-opus-4-5-20251101-thinking-64k-high-effort66.45
10gpt-5.4-xhigh65.901
11claude-opus-4-7-xhigh-effort61.409
12glm-5.260.725
13qwen3.7-max58.717
14gpt-5.2-2025-12-11-high58.427
15kimi-k2.6-thinking58.091
16claude-sonnet-4-6-thinking-auto-medium-effort56.975
17deepseek-v4-pro56.391
18gpt-5.2-codex56.035
19kimi-k2.7-code55.717
20claude-sonnet-5-xhigh-effort53.411
21grok-build-0.153.38
22qwen3.6-plus52.468
23minimax-m350.174
24grok-4.348.74
25deepseek-v4-flash46.872
26gpt-5.4-mini-xhigh43.875
27qwen3.6-27b42.746
28gpt-5.4-nano-xhigh39.133