LiveBench browser leaderboard
- Claude Fable 5.1 leads
- 100.0%
- models measured
- 43
- lead over #2
- +4.0 pts
- median
- 68.0%
Leader on LiveBench browser · read October 22
On October 22, 2026, 43 of our catalog's models had a LiveBench browser result, and Anthropic Claude Fable 5.1 led with 100.0%, ahead of Anthropic Claude Opus 4.8 at 96.0%.
Multi-step tasks in working web applications, done through a real browser; a model that cannot take screenshots scores 0. Each score is the model's mean task score for the category, 0 to 100, from a single run of the current release, shown as a share.
Role in the index
LiveBench browser carries 20% of the Agentic Index. A model missing it is scored on the other benchmarks of that index. Read the methodology
| # | Model | Result | Cost per task | Output tokens per task | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Claude Fable 5.1Anthropic | 100.0% | $3.7462 | 18,605 | 84.2 | $50.00 | |
| 2 | Claude Opus 4.8Anthropic | 96.0% | $8.8717 | 52,703 | 65.0 | $25.00 | |
| 3 | Claude Fable 5Anthropic | 92.0% | $10.3016 | 25,561 | 79.3 | $50.00 | |
| 4 | Gemini 4 ArgonGoogle | 92.0% | $3.4339 | 15,322 | 81.0 | $10.00 | |
| 5 | GPT-6 SolOpenAI | 92.0% | $3.2838 | 7,523 | 76.0 | $10.00 | |
| 6 | GPT-5.6 SolOpenAI | 88.0% | $1.3382 | 7,727 | 73.5 | $20.00 | |
| 7 | GPT-6.1 SolOpenAI | 88.0% | $2.4373 | 4,431 | 82.4 | $10.00 | |
| 8 | MiMo-V2.6-ProXiaomi | 88.0% | $0.4567 | 33,173 | 65.4 | $0.87 | |
| 9 | Claude Opus 5Anthropic | 84.0% | $2.9653 | 17,216 | 83.0 | $25.00 | |
| 10 | Claude Sonnet 5.5Anthropic | 84.0% | $7.6588 | 40,564 | 87.3 | $10.00 | |
| 11 | DeepSeek V4.1 FlashDeepSeek | 84.0% | $0.3486 | 42,493 | 62.5 | $0.84 | |
| 12 | GPT-5.6 TerraOpenAI | 84.0% | $0.8472 | 11,017 | 63.9 | $12.00 | |
| 13 | GPT-6 AstraOpenAI | 84.0% | $5.9097 | 3,639 | 85.1 | $50.00 | |
| 14 | GLM 5.3 FlashZ.ai | 84.0% | $0.1769 | 17,113 | 57.7 | $0.50 | |
| 15 | Claude Haiku 5.5Anthropic | 80.0% | $0.46 | 42,653 | 70.7 | $0.50 | |
| 16 | Claude Opus 5.5Anthropic | 80.0% | $14.6698 | 34,270 | 88.8 | $20.00 | |
| 17 | Muse Spark 1.3Meta | 80.0% | $1.6898 | 21,499 | 57.4 | $4.25 | |
| 18 | DeepSeek V4.1 Flash 1017DeepSeek | 76.0% | $0.245 | 26,695 | 70.3 | $0.60 | |
| 19 | Claude Sonnet 5Anthropic | 72.0% | $6.2898 | 26,428 | 57.7 | $10.00 | |
| 20 | GPT-6 LunaOpenAI | 72.0% | $0.2382 | 28,295 | 57.3 | $0.50 | |
| 21 | Claude Sonnet 4.6Anthropic | 68.0% | $3.3753 | 22,502 | 57.3 | $15.00 | |
| 22 | GPT-5.6 LunaOpenAI | 68.0% | $0.0973 | 10,159 | 58.0 | $1.20 | |
| 23 | DeepSeek V4 Flash Vision ExpDeepSeek | 64.0% | $0.4317 | 53,627 | Not benchmarked | $1.32 | |
| 24 | Gemini 3.8 FlashGoogle | 64.0% | $1.2131 | 36,860 | 61.6 | $3.75 | |
| 25 | Grok 4.6xAI | 64.0% | $2.9212 | 16,144 | 65.7 | $6.00 | |
| 26 | Grok 4.7xAI | 60.0% | $4.7778 | 36,667 | 66.7 | $6.00 | |
| 27 | Gemini 3.7 FlashGoogle | 56.0% | $1.0693 | 27,807 | 58.0 | $3.75 | |
| 28 | GPT-5.5OpenAI | 48.0% | $5.62 | 79,713 | 71.3 | $30.00 | |
| 29 | Qwen3.8 FlashQwen | 48.0% | $0.224 | 35,680 | Not benchmarked | $0.47 | |
| 30 | Grok 4.5xAI | 44.0% | $2.9177 | 12,613 | 58.6 | $6.00 | |
| 31 | Claude Haiku 4.5Anthropic | 40.0% | $1.5018 | 24,460 | 35.6 | $5.00 | |
| 32 | Muse Spark 1.2Meta | 36.0% | $2.4945 | 36,088 | 55.7 | $4.25 | |
| 33 | Muse Spark OpenMeta | 36.0% | $2.6434 | 23,505 | 47.4 | $4.25 | |
| 34 | Qwen3.8 27BQwen | 32.0% | $1.2903 | 49,468 | 74.2 | $2.55 | |
| 35 | Qwen3.8 Max (0902)Qwen | 32.0% | $1.5426 | 45,240 | 62.4 | $6.00 | |
| 36 | Step 5 PreviewStepFun | 24.0% | $2.7917 | 107,108 | 61.1 | $2.70 | |
| 37 | Kimi K3Moonshot AI | 16.0% | $0.382 | 5,917 | 62.3 | $15.00 | |
| 38 | T InklingThinking Machines | 16.0% | $2.3962 | 28,102 | 31.5 | $4.05 | |
| 39 | MiniMax M3MiniMax | 8.0% | $0.4292 | 29,095 | 39.2 | $1.20 | |
| 40 | Mistral Large 4Mistral AI | 8.0% | $0.4506 | 6,551 | 56.4 | $4.18 | |
| 41 | Gemma 4 31BGoogle | 4.0% | $0.2541 | 100,134 | 39.0 | $0.40 | |
| 42 | Nemotron 3.5 Super VLNVIDIA | 4.0% | $0.3509 | 32,614 | Not benchmarked | $0.50 | |
| 43 | GLM 5.3Z.ai | 0.0% | Not published | Not published | 70.6 | $4.40 |
Only models in our catalog are listed, and each appears once with its best published run. 43 rows were read from the source on October 22, 2026, and rows that match no catalog model are left out. Cost per task is what LiveBench reports the run spent on the category, divided by the number of tasks in it, in US dollars; LiveBench defines it the same way. Every other column is a figure the board publishes for the same run.
| Model | Result | Data Aggregation | Sequential Planning | Analysis-Driven Action | Structural Integrity | Constraint Adherence |
|---|---|---|---|---|---|---|
| Anthropic Claude Fable 5.1 | 100.0% | 100.0% | 100.0% | 100.0% | 100.0% | 100.0% |
| Anthropic Claude Opus 4.8 | 96.0% | 85.7% | 100.0% | 100.0% | 100.0% | 100.0% |
| Anthropic Claude Fable 5 | 92.0% | 71.4% | 100.0% | 100.0% | 100.0% | 100.0% |
| Google Gemini 4 Argon | 92.0% | 85.7% | 100.0% | 83.3% | 100.0% | 100.0% |
| OpenAI GPT-6 Sol | 92.0% | 85.7% | 100.0% | 83.3% | 100.0% | 100.0% |
| OpenAI GPT-5.6 Sol | 88.0% | 86.0% | 100.0% | 83.0% | 80.0% | 100.0% |
| OpenAI GPT-6.1 Sol | 88.0% | 100.0% | 100.0% | 83.3% | 80.0% | 66.7% |
| Xiaomi MiMo-V2.6-Pro | 88.0% | 71.4% | 75.0% | 100.0% | 100.0% | 100.0% |
| Anthropic Claude Opus 5 | 84.0% | 57.1% | 100.0% | 100.0% | 80.0% | 100.0% |
| Anthropic Claude Sonnet 5.5 | 84.0% | 71.4% | 50.0% | 100.0% | 100.0% | 100.0% |
| DeepSeek V4.1 Flash | 84.0% | 71.4% | 100.0% | 100.0% | 60.0% | 100.0% |
| OpenAI GPT-5.6 Terra | 84.0% | 71.4% | 100.0% | 83.3% | 80.0% | 100.0% |
| OpenAI GPT-6 Astra | 84.0% | 71.4% | 100.0% | 83.3% | 100.0% | 66.7% |
| Z.ai GLM 5.3 Flash | 84.0% | 57.1% | 100.0% | 100.0% | 80.0% | 100.0% |
| Anthropic Claude Haiku 5.5 | 80.0% | 57.1% | 75.0% | 100.0% | 100.0% | 66.7% |
| Anthropic Claude Opus 5.5 | 80.0% | 71.4% | 50.0% | 83.3% | 100.0% | 100.0% |
| Meta Muse Spark 1.3 | 80.0% | 71.4% | 50.0% | 100.0% | 80.0% | 100.0% |
| DeepSeek V4.1 Flash 1017 | 76.0% | 71.4% | 100.0% | 66.7% | 80.0% | 66.7% |
| Anthropic Claude Sonnet 5 | 72.0% | 71.4% | 50.0% | 83.3% | 80.0% | 66.7% |
| OpenAI GPT-6 Luna | 72.0% | 71.4% | 75.0% | 83.3% | 40.0% | 100.0% |
| Anthropic Claude Sonnet 4.6 | 68.0% | 57.1% | 100.0% | 83.3% | 40.0% | 66.7% |
| OpenAI GPT-5.6 Luna | 68.0% | 71.4% | 75.0% | 66.7% | 40.0% | 100.0% |
| DeepSeek V4 Flash Vision Exp | 64.0% | 71.4% | 75.0% | 50.0% | 60.0% | 66.7% |
| Google Gemini 3.8 Flash | 64.0% | 42.9% | 25.0% | 83.3% | 80.0% | 100.0% |
| SpaceXAI Grok 4.6 | 64.0% | 42.9% | 25.0% | 83.3% | 80.0% | 100.0% |
| SpaceXAI Grok 4.7 | 60.0% | 85.7% | 25.0% | 16.7% | 100.0% | 66.7% |
| Google Gemini 3.7 Flash | 56.0% | 42.9% | 50.0% | 33.3% | 80.0% | 100.0% |
| OpenAI GPT-5.5 | 48.0% | 57.1% | 100.0% | 0.0% | 20.0% | 100.0% |
| Qwen3.8 Flash | 48.0% | 42.9% | 25.0% | 50.0% | 60.0% | 66.7% |
| SpaceXAI Grok 4.5 | 44.0% | 57.1% | 25.0% | 0.0% | 60.0% | 100.0% |
| Anthropic Claude Haiku 4.5 | 40.0% | 14.3% | 50.0% | 50.0% | 40.0% | 66.7% |
| Meta Muse Spark 1.2 | 36.0% | 57.1% | 0.0% | 16.7% | 60.0% | 33.3% |
| Meta Muse Spark Open | 36.0% | 57.1% | 25.0% | 16.7% | 20.0% | 66.7% |
| Qwen3.8 27B | 32.0% | 57.1% | 0.0% | 0.0% | 80.0% | 0.0% |
| Qwen3.8 Max (0902) | 32.0% | 42.9% | 0.0% | 50.0% | 20.0% | 33.3% |
| StepFun Step 5 Preview | 24.0% | 28.6% | 25.0% | 0.0% | 40.0% | 33.3% |
| MoonshotAI Kimi K3 | 16.0% | 28.6% | 25.0% | 0.0% | 20.0% | 0.0% |
| Thinking Machines Inkling | 16.0% | 42.9% | 0.0% | 16.7% | 0.0% | 0.0% |
| MiniMax M3 | 8.0% | 28.6% | 0.0% | 0.0% | 0.0% | 0.0% |
| Mistral Large 4 | 8.0% | 28.6% | 0.0% | 0.0% | 0.0% | 0.0% |
| Google Gemma 4 31B | 4.0% | 14.3% | 0.0% | 0.0% | 0.0% | 0.0% |
| NVIDIA Nemotron 3.5 Super VL | 4.0% | 14.3% | 0.0% | 0.0% | 0.0% | 0.0% |
| Z.ai GLM 5.3 | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
The parts as LiveBench publishes them, for the same run as the result.