Inference
LiveBench

LiveBench Data Analysis

Mean score over joining tables, reformatting them and ordering events in them, using datasets published too recently to have been trained on.

ModelsModels with a result.55
Top resultBest result on this test.83.0%GPT-6 Astra Max
Top-3 spreadPoints from first to third.2.4 pts
UpdatedDate the source changed the results.25 Jun 2026

Result and cost

020406080100$0.003$0.01$0.03$0.1$0.3$1$3RESULTUS DOLLARS PER TASK · LOG SCALE

Results

55 results
#ModelResultThe score from the source.CostUS dollars to run one task.
01GPT-6 Astra Max83.0%$0.699
02GPT-5.5xhigh effort81.6%$0.405
03Claude Fable 5max effort80.5%$1.3
04Claude Opus 5.5max effort80.3%$0.815
05Claude Fable 5.1max effort80.3%$1.33
06Smaug Agentic79.9%$0.331
07GPT-5.6 Sol Max79.8%$0.353
08Muse Spark 1.3xhigh effort79.6%$0.169
09DeepSeek V4 Flash Vision Exp79.5%$0.04
10GPT-5.4xhigh effort79.3%$0.282
11GPT-5.6 Terra Max79.3%$0.278
12DeepSeek V4.1 Flash Max79.3%$0.036
13Smaug Flash79.1%<$0.01
14Smaug Mini78.8%$0.027
15Qwen3.8 Max78.4%$0.219
16Claude Opus 4.7xhigh effort78.3%$0.498
17GPT-5.2-Codex78.2%$0.199
18GPT-5.2high effort78.2%$0.166
19GPT-5.6 Luna Max78.0%$0.134
20Claude Sonnet 4.6 Thinking Automedium effort77.9%$0.326
21DeepSeek V4 Pro 042376.9%$0.043
22Grok 4.7xhigh effort76.9%$0.488
23Qwen3.8-27B76.6%$0.02
24Muse Spark 1.2xhigh effort76.5%$0.109
25GLM-5.3-Flash76.4%$0.024
26MiniMax M376.2%$0.029
27Ox Alpha Max75.8%$0
28Union Alpha74.6%$0
29Claude Opus 5max effort74.6%$0.746
30Claude Opus 4.5 Thinking 64Khigh effort74.4%$0.518
31Qwen3.8-Flash-Next74.2%$0.044
32Grok 4.673.9%$0.12
33GLM-5.273.7%$0.113
34DeepSeek V4 Flash 042373.7%$0.013
35Grok 4.573.0%$0.115
36Inklingxhigh effort72.8%$0.17
37Muse Spark 1.1xhigh effort72.5%$0.087
38Qwen3.7 Max71.8%$0.121
39Claude Sonnet 5xhigh effort71.7%$0.465
40Grok Build 0.170.8%<$0.01
41GPT-5.4 minixhigh effort70.8%$0.197
42Qwen3.6-27B70.4%$0.088
43GLM-5.370.2%$0.373
44Qwen3.6 Plus69.9%$0.077
45Claude Opus 4.6 Thinking Autohigh effort69.9%$0.362
46Gemini 3.7 Flashhigh effort68.0%$0.167
47GPT-5.4 nanoxhigh effort67.6%$0.045
48Claude Opus 4.8max effort66.0%$0.646
49Kimi K2.6 Thinking65.1%$0.125
50Gemini 3.5 Flashhigh effort64.9%$0.246
51Gemini 3.6 Flashhigh effort63.0%$0.191
52Kimi K2.7 Code62.7%$0.089
53Grok 4.355.8%$0.035
54Gemini 3.8 Flashhigh effort54.0%$0.238
55Gemini 3.5 Flash-Litehigh effort53.2%$0.045