Inference
EN
LiveBench

LiveBench Mathematics

Mean score over synthetic AMPS problems, integrals, recent competition questions and olympiad problem rearrangement. Not in LiveBench 2.0.

ModelsModels with a result.66
Top resultBest result on this benchmark.97.1%Claude Opus 5.5
Top-3 spreadPoints from first to third.0.2 pts
UpdatedDate the source changed the results.25 Jun 2026

Result and cost

65 of 66 results

Results

66 results
#ModelResultThe score from the source.CostUS dollars to run one task.
01Claude Opus 5.5max effort97.1%$0.221
02Claude Fable 5.1max effort97.0%$0.589
03GPT-6.1 Solmax effort96.8%$0.049
03GPT-6 Astramax effort96.8%$0.248
05Muse Wondermax effort96.5%$0.219
06GPT-6 Solmax effort96.4%$0.085
07GPT-5.6 Solmax effort96.2%$0.138
08Claude Sonnet 5.5max effort96.1%$0.168
09Claude Fable 5max effort96.0%$0.751
10Muse Spark 1.3xhigh effort95.9%$0.104
10GPT-5.5xhigh effort95.9%$0.196
12Claude Opus 5max effort95.7%$0.257
12Grok 4.7xhigh effort95.7%$0.335
14DeepSeek V4 Pro 081395.1%$0.031
15GPT-5.6 Terramax effort94.9%$0.132
16Claude Opus 4.8max effort94.3%$0.512
17GPT-5.4xhigh effort94.1%$0.165
18Mistral Large 4high effort93.6%$0.05
19Gemini 3.7 Flashhigh effort93.5%$0.074
20DeepSeek V4.1 Flash 1017max effort93.3%$0.022
21GPT-5.2high effort93.2%$0.109
22Claude Sonnet 5xhigh effort92.9%$0.295
22Claude Opus 4.7xhigh effort92.9%$0.299
24Grok 4.692.6%$0.064
25Gemini 3.8 Flashhigh effort91.6%$0.128
26Qwen3.8 Max91.3%$0.103
27Muse Spark 1.2xhigh effort91.2%$0.063
28Gemini 3.1 Pro Previewhigh effort91.0%$0.171
28GPT-5.4 nanoxhigh effort91.0%$0.026
30Grok 4.590.8%$0.066
31DeepSeek V4 Pro 042390.7%$0.035
32Claude Opus 4.5 Thinking 64Khigh effort90.4%$0.404
33Smaug Flash90.1%$0.0054
34GLM-5.289.8%$0.096
35Smaug Mini89.7%$0.015
36Claude Opus 4.6 Thinking Autohigh effort89.3%$0.254
37GPT-6 Lunamax effort89.1%$0.014
38GPT-5.2-Codex88.8%$0.108
39Inklingxhigh effort88.4%$0.106
40Gemini 3.5 Flashhigh effort88.2%$0.138
41GLM-5.387.9%$0.266
42DeepSeek V4 Flash Vision Exp87.8%$0.029
43GPT-5.6 Lunamax effort87.2%$0.062
44Muse Spark 1.1xhigh effort87.1%$0.046
45Claude Sonnet 4.6 Thinking Automedium effort87.0%$0.233
46DeepSeek V4 Flash 073186.8%$0.0047
47Claude Haiku 5.5max effort86.4%$0.016
47Gemini 3.6 Flashhigh effort86.4%$0.092
49Qwen3.8-27B86.2%$0.015
50Qwen3.8-Flash-Next85.8%$0.022
51Qwen3.7 Max85.2%$0.127
52Kimi K384.4%$0.21
53Grok 4.384.3%$0.034
53Kimi K2.6 Thinking84.3%$0.095
55Smaug Agentic83.9%$0.172
56Qwen3.6 Plus83.7%$0.106
57GLM-5.3-Flash81.2%$0.022
58Qwen3.6-27B79.9%$0.123
59DeepSeek V4 Flash 042379.6%$0.01
59Kimi K2.7 Code79.6%$0.048
61GPT-5.4 minixhigh effort78.5%$0.147
62Grok Build 0.178.4%$0.0016
63Ox Alpha Max77.5%$0
64MiniMax M376.9%$0.016
65Gemini 3.5 Flash-Litehigh effort73.7%$0.028
66Nemotron 3.5 Lightning 30B A3B66.5%$0.01