Inference
LiveBench

LiveBench Coding

Mean score over writing a solution to a fresh competition problem and completing one whose body was removed.

ModelsModels with a result.55
Top resultBest result on this test.89.3%Claude Opus 5.5
Top-3 spreadPoints from first to third.3.3 pts
UpdatedDate the source changed the results.25 Jun 2026

Result and cost

020406080100$0.0003$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 effort89.3%$0.303
02Claude Fable 5.1max effort86.4%$0.401
03Claude Fable 5max effort86.0%$0.212
04GPT-5.6 Sol Max83.9%$0.107
05GPT-5.2-Codex83.6%$0.029
06GPT-5.6 Luna Max82.9%$0.052
07Smaug Agentic82.5%$0.058
08GPT-5.5xhigh effort82.1%$0.154
08Union Alpha82.1%$0
10Claude Opus 4.7xhigh effort82.1%$0.029
11Claude Opus 4.8max effort81.8%$0.123
12Claude Opus 5max effort81.4%$0.033
13Muse Spark 1.3xhigh effort81.1%$0.059
14Claude Sonnet 5xhigh effort80.7%$0.035
15GPT-6 Astra Max80.4%$0.296
16DeepSeek V4.1 Flash Max80.0%$0.013
17GLM-5.279.7%$0.051
17Claude Opus 4.5 Thinking 64Khigh effort79.7%$0.174
19Claude Sonnet 4.6 Thinking Automedium effort79.3%<$0.01
20GLM-5.379.0%$0.17
20GLM-5.3-Flash79.0%<$0.01
22Gemini 3.7 Flashhigh effort78.9%$0.051
23Kimi K2.6 Thinking78.6%$0.028
24GPT-5.6 Terra Max78.2%$0.169
25Gemini 3.5 Flashhigh effort78.2%$0.067
25Qwen3.6 Plus78.2%$0.017
25Claude Opus 4.6 Thinking Autohigh effort78.2%$0.03
28Gemini 3.6 Flashhigh effort77.9%$0.076
29Muse Spark 1.2xhigh effort77.5%$0.051
29GPT-5.4xhigh effort77.5%$0.267
31Grok 4.7xhigh effort77.2%$0.302
31Muse Spark 1.1xhigh effort77.2%$0.024
33Smaug Flash77.1%<$0.01
34Grok 4.676.8%$0.021
35GPT-5.2high effort76.1%$0.029
35Gemini 3.5 Flash-Litehigh effort76.1%$0.01
37Ox Alpha Max75.8%$0
38Qwen3.8-27B75.7%$0.019
39Smaug Mini75.4%$0.01
40Qwen3.7 Max74.2%$0.027
41Kimi K2.7 Code74.0%$0.014
42DeepSeek V4 Pro 042373.6%<$0.01
43Qwen3.8 Max72.9%$0.113
44Qwen3.8-Flash-Next72.6%$0.028
45Gemini 3.8 Flashhigh effort72.5%$0.185
46DeepSeek V4 Flash 042372.1%<$0.01
47Qwen3.6-27B71.8%$0.02
48GPT-5.4 minixhigh effort71.6%$0.222
49Inklingxhigh effort71.0%$0.053
50GPT-5.4 nanoxhigh effort70.8%$0.015
51Grok 4.369.9%$0.011
52Grok 4.568.6%$0.021
53MiniMax M368.2%<$0.01
53DeepSeek V4 Flash Vision Exp68.2%$0.035
55Grok Build 0.165.4%<$0.01