LiveBench data analysis leaderboard
- Gemini 4 Argon leads
- 85.2%
- models measured
- 43
- lead over #2
- +2.2 pts
- median
- 76.4%
Leader on LiveBench data analysis · read October 21
On October 21, 2026, 43 of our catalog's models had a LiveBench data analysis result, and Google Gemini 4 Argon led with 85.2%.
Short data-analysis questions on small tables (runs of consecutive growth, joining one table's columns onto another's, converting a table between formats), each with a single correct answer graded against ground truth. 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
We collect this benchmark and show it as supporting evidence. It is not part of the Vibe Coding Index, because it measures something other than coding and the index is about coding. Read the methodology
| # | Model | Result | Cost per task | Output tokens per task | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Gemini 4 ArgonGoogle | 85.2% | $0.2843 | 26,732 | 80.5 | $10.00 | |
| 2 | GPT-6 AstraOpenAI | 83.0% | $0.6989 | 13,240 | 88.7 | $50.00 | |
| 3 | GPT-6.1 SolOpenAI | 82.7% | $0.1454 | 13,804 | 88.7 | $10.00 | |
| 4 | GPT-5.5OpenAI | 81.6% | $0.4055 | 13,299 | 67.8 | $30.00 | |
| 5 | GPT-6 SolOpenAI | 81.2% | $0.1989 | 19,154 | 82.8 | $10.00 | |
| 6 | Claude Fable 5Anthropic | 80.5% | $1.3007 | 25,207 | 82.1 | $50.00 | |
| 7 | Claude Fable 5.1Anthropic | 80.3% | $1.3331 | 25,856 | 89.1 | $50.00 | |
| 8 | Claude Opus 5.5Anthropic | 80.3% | $0.8146 | 39,924 | 92.2 | $20.00 | |
| 9 | GPT-5.6 SolOpenAI | 79.8% | $0.2381 | 11,094 | 76.7 | $20.00 | |
| 10 | Muse Spark 1.3Meta | 79.6% | $0.1694 | 38,795 | 79.7 | $4.25 | |
| 11 | DeepSeek V4 Flash Vision ExpDeepSeek | 79.5% | $0.079 | 59,825 | 37.6 | $0.6468 | |
| 12 | DeepSeek V4.1 Flash 1017DeepSeek | 79.5% | $0.0332 | 26,704 | 83.3 | $0.60 | |
| 13 | DeepSeek V4.1 FlashDeepSeek | 79.3% | $0.0711 | 58,326 | 77.2 | $0.525 | |
| 14 | GPT-5.6 TerraOpenAI | 79.3% | $0.2222 | 17,623 | 69.0 | $12.00 | |
| 15 | Kimi K3Moonshot AI | 78.7% | $0.3524 | 22,775 | 61.8 | $11.936 | |
| 16 | Qwen3.8 Max (0902)Qwen | 78.4% | $0.2189 | 34,658 | 63.5 | $6.00 | |
| 17 | GPT-5.6 LunaOpenAI | 78.0% | $0.0267 | 21,599 | 61.5 | $1.20 | |
| 18 | Grok 4.7xAI | 76.9% | $0.488 | 80,144 | 75.0 | $6.00 | |
| 19 | Qwen3.8 27BQwen | 76.6% | $0.089 | 29,045 | 73.6 | $1.35 | |
| 20 | Muse Spark 1.2Meta | 76.5% | $0.1094 | 25,752 | 65.1 | $4.25 | |
| 21 | Mistral Large 4Mistral AI | 76.5% | $0.1337 | 62,776 | 55.3 | $4.18 | |
| 22 | GLM 5.3 FlashZ.ai | 76.4% | $0.0243 | 47,359 | 71.0 | $0.50 | |
| 23 | MiniMax M3MiniMax | 76.2% | $0.0288 | 23,291 | 57.8 | $1.20 | |
| 24 | Claude Sonnet 4.6Anthropic | 76.1% | $0.4169 | 27,034 | 57.8 | $15.00 | |
| 25 | Grok 4.6xAI | 75.3% | $0.123 | 19,229 | 76.4 | $6.00 | |
| 26 | Claude Opus 5Anthropic | 74.6% | $0.7455 | 29,014 | 81.8 | $25.00 | |
| 27 | Qwen3.8 FlashQwen | 74.2% | $0.0435 | 90,692 | 87.4 | $0.47 | |
| 28 | GPT-6 LunaOpenAI | 73.4% | $0.0166 | 32,473 | 71.4 | $0.50 | |
| 29 | Grok 4.5xAI | 73.0% | $0.1139 | 18,701 | 75.5 | $6.00 | |
| 30 | T InklingThinking Machines | 72.8% | $0.1446 | 34,889 | 49.7 | $4.05 | |
| 31 | MiMo-V2.6-ProXiaomi | 72.3% | $0.0386 | 41,652 | 72.6 | $0.87 | |
| 32 | Claude Sonnet 5Anthropic | 71.7% | $0.3101 | 30,244 | 63.5 | $10.00 | |
| 33 | GLM 5.3Z.ai | 70.2% | $0.3731 | 83,548 | 73.1 | $1.40 | |
| 34 | Gemini 3.7 FlashGoogle | 68.0% | $0.1667 | 43,360 | 69.9 | $3.75 | |
| 35 | Claude Opus 4.8Anthropic | 66.0% | $0.6456 | 25,286 | 66.0 | $25.00 | |
| 36 | Nemotron 3.5 Super VLNVIDIA | 64.6% | $0.0194 | 31,654 | 65.5 | $0.50 | |
| 37 | Step 5 PreviewStepFun | 63.5% | $0.0639 | 22,293 | 71.4 | $2.70 | |
| 38 | Muse Spark OpenMeta | 60.2% | $0.0449 | 9,999 | 62.1 | $4.25 | |
| 39 | Claude Sonnet 5.5Anthropic | 59.5% | $0.5306 | 52,256 | 88.9 | $10.00 | |
| 40 | Claude Haiku 4.5Anthropic | 59.3% | Not published | Not published | 49.8 | $5.00 | |
| 41 | Gemini 3.8 FlashGoogle | 54.0% | $0.2377 | 62,326 | 70.5 | $4.125 | |
| 42 | Gemma 4 31BGoogle | 53.3% | $0.0046 | 9,487 | 38.3 | $0.40 | |
| 43 | Claude Haiku 5.5Anthropic | 53.1% | $0.0315 | 61,936 | 75.9 | $0.50 |
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 21, 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 | Consecutive events | Table join | Table reformat |
|---|---|---|---|---|
| Google Gemini 4 Argon | 85.2% | 91.4% | 69.9% | 94.1% |
| OpenAI GPT-6 Astra | 83.0% | 90.6% | 58.4% | 100.0% |
| OpenAI GPT-6.1 Sol | 82.7% | 90.6% | 57.4% | 100.0% |
| OpenAI GPT-5.5 | 81.6% | 88.8% | 55.9% | 100.0% |
| OpenAI GPT-6 Sol | 81.2% | 89.5% | 54.0% | 100.0% |
| Anthropic Claude Fable 5 | 80.5% | 91.4% | 56.1% | 94.1% |
| Anthropic Claude Fable 5.1 | 80.3% | 90.6% | 56.2% | 94.1% |
| Anthropic Claude Opus 5.5 | 80.3% | 90.6% | 52.3% | 98.0% |
| OpenAI GPT-5.6 Sol | 79.8% | 90.2% | 49.3% | 100.0% |
| Meta Muse Spark 1.3 | 79.6% | 90.5% | 48.2% | 100.0% |
| DeepSeek V4 Flash Vision Exp | 79.5% | 90.0% | 50.3% | 98.0% |
| DeepSeek V4.1 Flash 1017 | 79.5% | 90.3% | 48.2% | 100.0% |
| DeepSeek V4.1 Flash | 79.3% | 88.7% | 49.0% | 100.0% |
| OpenAI GPT-5.6 Terra | 79.3% | 87.7% | 50.2% | 100.0% |
| MoonshotAI Kimi K3 | 78.7% | 89.8% | 48.4% | 98.0% |
| Qwen3.8 Max (0902) | 78.4% | 87.1% | 50.1% | 98.0% |
| OpenAI GPT-5.6 Luna | 78.0% | 86.5% | 47.6% | 100.0% |
| SpaceXAI Grok 4.7 | 76.9% | 85.7% | 44.9% | 100.0% |
| Qwen3.8 27B | 76.6% | 84.9% | 48.8% | 96.1% |
| Meta Muse Spark 1.2 | 76.5% | 82.5% | 46.8% | 100.0% |
| Mistral Large 4 | 76.5% | 80.5% | 49.1% | 100.0% |
| Z.ai GLM 5.3 Flash | 76.4% | 84.3% | 48.8% | 96.1% |
| MiniMax M3 | 76.2% | 84.6% | 45.8% | 98.0% |
| Anthropic Claude Sonnet 4.6 | 76.1% | 82.9% | 49.2% | 96.1% |
| SpaceXAI Grok 4.6 | 75.3% | 84.6% | 43.1% | 98.0% |
| Anthropic Claude Opus 5 | 74.6% | 77.6% | 52.0% | 94.1% |
| Qwen3.8 Flash | 74.2% | 75.6% | 51.0% | 96.1% |
| OpenAI GPT-6 Luna | 73.4% | 70.5% | 49.6% | 100.0% |
| SpaceXAI Grok 4.5 | 73.0% | 75.5% | 43.6% | 100.0% |
| Thinking Machines Inkling | 72.8% | 69.9% | 48.5% | 100.0% |
| Xiaomi MiMo-V2.6-Pro | 72.3% | 70.5% | 48.3% | 98.0% |
| Anthropic Claude Sonnet 5 | 71.7% | 74.7% | 42.4% | 98.0% |
| Z.ai GLM 5.3 | 70.2% | 72.3% | 42.3% | 96.1% |
| Google Gemini 3.7 Flash | 68.0% | 60.4% | 45.5% | 98.0% |
| Anthropic Claude Opus 4.8 | 66.0% | 52.9% | 51.1% | 94.1% |
| NVIDIA Nemotron 3.5 Super VL | 64.6% | 49.6% | 44.3% | 100.0% |
| StepFun Step 5 Preview | 63.5% | 42.4% | 50.1% | 98.0% |
| Meta Muse Spark Open | 60.2% | 36.0% | 44.5% | 100.0% |
| Anthropic Claude Sonnet 5.5 | 59.5% | 40.6% | 39.8% | 98.0% |
| Anthropic Claude Haiku 4.5 | 59.3% | 39.2% | 42.6% | 96.1% |
| Google Gemini 3.8 Flash | 54.0% | 20.6% | 45.3% | 96.1% |
| Google Gemma 4 31B | 53.3% | 13.6% | 46.3% | 100.0% |
| Anthropic Claude Haiku 5.5 | 53.1% | 16.4% | 44.8% | 98.0% |
The parts as LiveBench publishes them, for the same run as the result.