Inference
EN
LiveBench

LiveBench Design

Make or restyle Word documents, PDFs and slide decks to a brief. Visual and structural checks grade the file.

ModelsModels with a result.52
Top resultBest result on this benchmark.86.9%Claude Opus 5.5
Top-3 spreadPoints from first to third.2.2 pts
UpdatedDate the source changed the results.14 Oct 2026

Result and cost

52 of 52 results

Results

52 results
#ModelResultThe score from the source.CostUS dollars to run one task.
01Claude Opus 5.5max effort86.9%$5.75
02Claude Sonnet 5.5max effort85.0%$5.02
03Claude Fable 5max effort84.7%$8.55
04Claude Fable 5.1max effort83.4%$7.38
05Claude Haiku 5.5max effort83.1%$0.663
06Claude Opus 5max effort83.0%$5.45
07GPT-6 Lunamax effort82.5%$0.031
08GPT-6.1 Terramax effort82.3%$0.27
09GPT-6 Solmax effort82.2%$0.425
10GPT-6.1 Solmax effort82.1%$0.371
11GPT-6 Astramax effort82.0%$2.88
11Claude Sonnet 5max effort82.0%$5.47
13GPT-5.6 Solmax effort81.6%$1.71
14Grok 4.6xhigh effort79.8%$1.27
15DeepSeek V4.1 Flash 1017max effort79.2%$0.175
16Claude Opus 4.8max effort78.8%$6.45
17Qwen3.8 Maxmax effort78.6%$0.596
18MiMo-V2.6-Prohigh effort78.5%$0.093
19Grok 4.5xhigh effort78.1%$0.388
20Grok 4.7xhigh effort78.0%$3.37
21GLM-5.3max effort77.3%$0.956
22Kimi K3max effort77.2%$1.00
22GLM-5.3-Flashmax effort77.2%$0.063
24Step 5 Previewmax effort76.9%$0.746
25Mistral Large 4.0high effort76.3%$0.206
26GPT-5.6 Terramax effort76.0%$0.62
26GPT-5.6 Lunamax effort76.0%$0.09
28Qwen3.8-27Bmax effort75.9%$0.384
29Muse Wondermax effort75.1%$0.687
30Muse Spark 1.3max effort74.8%$0.829
31MiniMax M3.1 Flashmax effort74.0%$0.158
32Gemini 4 Argonhigh effort73.7%$1.16
33Gemini 3.7 Flashhigh effort73.2%$0.623
34Qwen3.8-Flash-Nextmax effort72.8%$0.128
35Gemini 3.8 Flashhigh effort72.7%$0.98
36Claude Sonnet 4.6max effort71.8%$2.14
37DeepSeek V4 Flash 0731max effort71.7%$0.101
38DeepSeek V4 Flash Vision Expmax effort70.7%$0.059
39GPT-5.5xhigh effort70.3%$1.19
40Beam 1.0max effort67.9%$0.375
41Muse Spark 1.2xhigh effort67.1%$0.856
42MiniMax M3max effort66.2%$0.208
43Mistral Large 4high effort65.6%$0.389
44Smaug Nemotron 3.5 Nvfp4max effort62.7%$0.028
45Gemini 3.1 Pro Previewhigh effort61.6%$0.596
46Nemotron Super VL-3.5max effort60.7%$0.056
47Muse Spark Openmax effort60.3%$0.356
48Claude Haiku 4.5max effortAbove fitFar above this model's index.59.9%$0.156
49Gemma 4 31Bmax effort58.4%$0.015
50Inkling58.3%$0.286
51Nemotron Super VL-3.5 Nvfp4max effort57.1%$0.038
52Nemotron 3.5 Lightning 30B A3Bmax effortAbove fitFar above this model's index.43.9%$0.03