All tests
Data analysis
LiveBench data analysis
Working with tables and structured information.
Published by LiveBench under Apache-2.0. Compare it with other tests
Top score83.0%GPT-6 Astra (max), OpenAI
Models tested63from 11 makers
Median score76.2%Half the models score above it
Scale0 to 100%The share of tasks done right
Top 15
Each model at its best configuration, coloured by maker.
The best model from each maker
Each maker’s highest-scoring model on this test, and its rank.
| Maker | Best model | Score | Rank |
|---|---|---|---|
 OpenAI | GPT-6 Astra (max) | 83.0% | #1 of 63 |
 Anthropic | Claude Fable 5 (max effort) | 80.5% | #5 of 63 |
 DeepSeek | DeepSeek V4 Flash Vision Exp | 79.5% | #11 of 63 |
 Moonshot AI | Kimi K3 | 78.7% | #19 of 63 |
 Google | Gemini 3.1 Pro | 78.5% | #21 of 63 |
 Alibaba | Qwen3.8 Max | 78.4% | #22 of 63 |
 xAI | Grok 4.7 (xhigh) | 76.9% | #28 of 63 |
 Z.ai | GLM-5.3-Flash | 76.4% | #31 of 63 |
 MiniMax | MiniMax-M3 | 76.2% | #32 of 63 |
 NVIDIA | Nemotron 3 Ultra 550B A55B | 54.5% | #61 of 63 |
Every result
63 models, each once at its best configuration, highest score first. Models with the same score share a rank.
| Rank | Model | Maker | Score |
|---|---|---|---|
| #1 | GPT-6 Astra (max) | OpenAI | 83.0% |
| #2 | GPT-6.1 Sol (max) | OpenAI | 82.7% |
| #3 | GPT-5.5 (xhigh) | OpenAI | 81.6% |
| #4 | GPT-6 Sol (max) | OpenAI | 81.2% |
| #5 | Claude Fable 5 (max effort) | Anthropic | 80.5% |
| #6 | Claude Opus 5.5 (max effort) | Anthropic | 80.3% |
| #7 | Claude Fable 5.1 (max effort) | Anthropic | 80.3% |
| #8 | Smaug Agentic |  Not reported | 79.9% |
| #9 | GPT-5.6 Sol (max) | OpenAI | 79.8% |
| #10 | Muse Spark 1.3 (xhigh) | Meta AI | 79.6% |
| #11 | DeepSeek V4 Flash Vision Exp | DeepSeek | 79.5% |
| #12 | DeepSeek V4 Flash 0731 | DeepSeek | 79.3% |
| #13 | GPT-5.4 (xhigh) | OpenAI | 79.3% |
| #13 | GPT-5.6 Terra (max) | OpenAI | 79.3% |
| #15 | DeepSeek V4.1 Flash (max) | DeepSeek | 79.3% |
| #16 | DeepSeek V4 Pro 0813 | DeepSeek | 79.2% |
| #17 | Smaug Flash | Not reported | 79.1% |
| #18 | Smaug Mini | Not reported | 78.8% |
| #19 | Kimi K3 | Moonshot AI | 78.7% |
| #20 | Claude Sonnet 5.5 (xhigh effort) | Anthropic | 78.6% |
| #21 | Gemini 3.1 Pro | Google | 78.5% |
| #22 | Qwen3.8 Max | Alibaba | 78.4% |
| #23 | Claude Opus 4.7 (xhigh effort) | Anthropic | 78.3% |
| #24 | GPT-5.2 Codex | OpenAI | 78.2% |
| #25 | GPT-5.2 (2025 12 11 high) | OpenAI | 78.2% |
How it is scored
Arithmetic mean of the 3 published task scores in this category. Only complete results are ranked. Exact model and effort configurations are preserved.
Results from LiveBench (Apache-2.0), as published. Model names link to their pages where the site can match them; makers come from the model’s page, or else from the publisher.