// ranked · September 2026

Best LLM for Data Analysis

Ranked by LiveBench’s data analysis category: table reformatting, column-type inference and join reasoning over real datasets.

OpenRouter + LiveBenchAll rankingsAll models

GPT-6 Astra leads on LiveBench data analysis at 83.0, ahead of GPT-5.5 (81.6) and GPT-6 Sol (81.2).

GPT-6 Astra vs GPT-5.5, head to head

Top 10

#ModelData analysis scoreBlended / 1MContextCapabilities
1GPT-6 Astra

openai/gpt-6-astra

83.0$20.001.1M
reasoningtoolsvision
2GPT-5.5

openai/gpt-5.5

81.6$11.251.1M
reasoningtoolsvision
3GPT-6 Sol

openai/gpt-6-sol

81.2$4.001.1M
reasoningtoolsvision
4Claude Fable 5

anthropic/claude-fable-5

80.5$20.001M
reasoningtoolsvision
5Claude Opus 5.5

anthropic/claude-opus-5.5

80.3$8.001M
reasoningtoolsvision
6Claude Fable 5.1

anthropic/claude-fable-5.1

80.3$20.001M
reasoningtoolsvision
7GPT-5.6 Sol

openai/gpt-5.6-sol

79.8$4.001.1M
reasoningtoolsvision
8Muse Spark 1.3

meta/muse-spark-1.3

79.6$2.001.0M
reasoningtoolsvision
9DeepSeek V4 Flash Vision Exp

deepseek/deepseek-v4-flash-vision-exp

79.5$0.3301.0M
reasoningtoolsvision
10DeepSeek V4 Flash 0731

deepseek/deepseek-v4-flash-0731

79.3$0.1901.3M
reasoningtools

45 more models qualify — see the full leaderboard.

FAQ

What is the best LLM for Data Analysis in September 2026?

GPT-6 Astra leads on LiveBench data analysis at 83.0, ahead of GPT-5.5 (81.6) and GPT-6 Sol (81.2).

What is the best value in the top 10 for data analysis?

DeepSeek V4 Flash 0731. Lowest measured cost per point among the leaders — $0.0049 per point for a score of 79.3.

What is the best under $1.00 / 1m for data analysis?

DeepSeek V4 Flash Vision Exp. Highest score at a blended list price of $1.00 per 1M tokens or less — 79.5 at $0.330.

How is this ranking produced?

Ranked by LiveBench’s data analysis category: table reformatting, column-type inference and join reasoning over real datasets.

Other rankings

How this ranking is produced

  • One metric, stated above — nothing here is weighted or scored by us. Scores come from LiveBench release 2026-06-25; models without a published run don't appear in score-based rankings.
  • Price, context and capabilities — live from OpenRouter, refreshed every 15 minutes.
  • Ties — scores less than a point apart are called a tie; effort settings alone move a LiveBench score by more than that.

A public benchmark is someone else's workload. Before committing, see LLM & agent evaluation.