Skip to content

Benchmark

LiveBench

Math benchmarks measure how well models solve maths problems. Scores are percentages.

Models measured
71
Scores recorded
87
Sources
2
IQ weight
2.39

Score distribution

72.4%
Mean
74.3%
Median
9.1%
Std dev
66.1%
10th percentile
79.5%
90th percentile
29.6%
Min
29.6%84.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.39 for chat models, measured over 70 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51GLM 5.2overall· LiveBench data73.4%
52QwQ 32B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim73.1%
53Claude Opus 4.5overall, 64k, high· LiveBench data72.9%
54Claude Haiku 5.5overall, xhigh· LiveBench data72.9%
55Inkling FP4overall, xhigh· LiveBench data72.9%
56Mistral Large 4overall, high· LiveBench data72.8%
57GPT-5.3 Codex· LLM Stats data72.8%
58DeepSeek V4 Prooverall· LiveBench data72.6%
59Gemini 3 Flash· LLM Stats data72.4%
60Kimi K2.6· LLM Stats data72.2%
61GPT 5.1high· LLM Stats data72%
62Kimi K2.7 Code· LLM Stats data71.9%
63GLM 5.3 Flashoverall· LiveBench data71.1%
64Kimi K2.6overall, thinking· LiveBench data70.9%
65Qwen3.6 Plus· LLM Stats data70.8%
66GPT 5.4 Nanooverall, xhigh· LiveBench data70.8%
67GLM 5.1· LLM Stats data70.2%
68GPT 5.4 Nano· LLM Stats data70.1%
69MiniMax M3· LLM Stats data70%
70Kimi K2.5· LLM Stats data69.1%
71Qwen3.6 Plusoverall· LiveBench data69%
72Kimi K2.7 Codeoverall· LiveBench data68.8%
73Nemotron 3 Ultraoverall· LiveBench data68.7%
74Grok Build 0.1overall· LiveBench data68.6%
75MiniMax M3overall· LiveBench data67.5%
76o1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim67%
77GPT-5.4 minioverall, xhigh· LiveBench data66.6%
78DeepSeek V4 Flashoverall· LiveBench data66.1%
79Qwen3.6 27Boverall· LiveBench data64.2%
80Gemini 3.5 Flash Liteoverall, high· LiveBench data63.8%
81Grok 4.3overall· LiveBench data63.3%
82O1 Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim52.3%
83Qwen2.5 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim52.3%
84Phi 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim47.6%
85Nemotron 3.5 Lightning 30B A3Boverall· LiveBench data45.8%
86Qwen2.5 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim35.9%
87Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim29.6%

Showing 51–87 of 87 results.

Data credits: Data by LLM Stats