Inference
LiveBench

LiveBench Mathematics

Mean score over synthetic AMPS problems, integrals, recent competition questions and olympiad problem rearrangement.

ModelsModels with a result.55
Top resultBest result on this test.97.1%Claude Opus 5.5
Top-3 spreadPoints from first to third.0.3 pts
UpdatedDate the source changed the results.25 Jun 2026

Result and cost

020406080100$0.001$0.003$0.01$0.03$0.1$0.3$1RESULTUS DOLLARS PER TASK · LOG SCALE

Results

55 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 Astra Max96.8%$0.248
04GPT-5.6 Sol Max96.2%$0.138
05Claude Fable 5max effort96.0%$0.751
06Muse Spark 1.3xhigh effort95.9%$0.104
07GPT-5.5xhigh effort95.9%$0.196
08Claude Opus 5max effort95.7%$0.257
09Grok 4.7xhigh effort95.7%$0.335
10Union Alpha95.3%$0
11GPT-5.6 Terra Max94.9%$0.132
12Claude Opus 4.8max effort94.3%$0.512
13GPT-5.4xhigh effort94.1%$0.165
14Gemini 3.7 Flashhigh effort93.5%$0.074
15DeepSeek V4.1 Flash Max93.3%$0.022
16GPT-5.2high effort93.2%$0.109
17Claude Sonnet 5xhigh effort92.9%$0.295
18DeepSeek V4 Pro 042392.9%$0.031
19Claude Opus 4.7xhigh effort92.9%$0.299
20Grok 4.692.6%$0.064
21Gemini 3.8 Flashhigh effort91.6%$0.128
22Qwen3.8 Max91.3%$0.103
23Muse Spark 1.2xhigh effort91.2%$0.063
24GPT-5.4 nanoxhigh effort91.0%$0.026
25Grok 4.590.8%$0.066
26Claude Opus 4.5 Thinking 64Khigh effort90.4%$0.404
27Smaug Flash90.1%<$0.01
28GLM-5.289.8%$0.096
29Smaug Mini89.7%$0.015
30Claude Opus 4.6 Thinking Autohigh effort89.3%$0.254
31GPT-5.2-Codex88.8%$0.108
32Inklingxhigh effort88.4%$0.106
33Gemini 3.5 Flashhigh effort88.2%$0.138
34GLM-5.387.9%$0.266
35DeepSeek V4 Flash Vision Exp87.8%$0.029
36GPT-5.6 Luna Max87.2%$0.062
37Muse Spark 1.1xhigh effort87.1%$0.046
38Claude Sonnet 4.6 Thinking Automedium effort87.0%$0.233
39Gemini 3.6 Flashhigh effort86.4%$0.092
40Qwen3.8-27B86.2%$0.015
41Qwen3.8-Flash-Next85.8%$0.022
42Qwen3.7 Max85.2%$0.127
43Grok 4.384.3%$0.034
44Kimi K2.6 Thinking84.3%$0.095
45Smaug Agentic83.9%$0.172
46Qwen3.6 Plus83.7%$0.106
47DeepSeek V4 Flash 042383.2%$0.01
48GLM-5.3-Flash81.2%$0.022
49Qwen3.6-27B79.9%$0.123
50Kimi K2.7 Code79.6%$0.048
51GPT-5.4 minixhigh effort78.5%$0.147
52Grok Build 0.178.4%<$0.01
53Ox Alpha Max77.5%$0
54MiniMax M376.9%$0.016
55Gemini 3.5 Flash-Litehigh effort73.7%$0.028