ModelCompare
Compare

AI model history

What was the top AI model in ?

Claude 3.7 Sonnet from Anthropic, scoring 17.7 on the Intelligence Index at $6 per million tokens. It had been on top for 5 weeks.

Top Intelligence Index score on each day. Click anywhere to go there.

Leaderboard

The top 10 on March 31, 2025

Ranked by the Artificial Analysis Intelligence Index, with each model's price on that day (per million tokens, 3 parts input to 1 part output).

  1. 1
    Claude 3.7 SonnetClosed
    Anthropic · released Feb 24, 2025 · 5 weeks old
    Intelligence 17.7
    $6 /M
  2. 2
    o1Closed
    OpenAI · released Dec 5, 2024 · 4 months old
    Intelligence 15.2
    $26 /M
  3. 3
    Gemini 2.5 Pro Preview (Mar' 25)Closed
    Google · released Mar 25, 2025 · 6 days old
    Intelligence 15
    No public price
  4. 4
    Grok 3 mini ReasoningClosed
    xAI · released Feb 19, 2025 · 6 weeks old
    Intelligence 14.6
    $0.35 /M
  5. 5
    o3-miniClosed
    OpenAI · released Jan 31, 2025 · 8 weeks old
    Intelligence 12.5
    $1.93 /M
  6. 6
    o1-proClosed
    OpenAI · released Mar 19, 2025 · 12 days old
    Intelligence 12.4
    $263 /M
  7. 7
    Grok 3Closed
    xAI · released Feb 19, 2025 · 6 weeks old
    Intelligence 12.1
    $6 /M
  8. 8
    Sonar Reasoning ProClosed
    Perplexity · released Jan 28, 2025 · 2 months old
    Intelligence 11.8
    No public price
  9. 9
    o1-previewClosed
    OpenAI · released Sep 12, 2024 · 7 months old
    Intelligence 11.4
    $29 /M
  10. 10
    DeepSeek R1 (Jan '25)Open weights
    DeepSeek · released Jan 20, 2025 · 2 months old
    Intelligence 11.4
    $2.50 /M

1.5 years later

What has happened since

#152Where Claude 3.7 Sonnet ranks today

361 models have come out since. Today's best is Claude Opus 5.5 at 57.6, 39.9 points higher.

60×Cheaper for the same score

Hy3-preview scores at least as well today for $0.10 per million tokens, against $6 for Claude 3.7 Sonnet back then.

4 monthsUntil it cost under $1

GPT-5 mini matched it for $0.69 per million tokens on Aug 7, 2025.

6 monthsUntil open weights caught up

GLM-4.6 from Z.ai matched it on Sep 30, 2025, free to download and run.

The race

Which lab held the top spot

OpenAI led for 2.5 years of the 4 years since ChatGPT launched. Anthropic has held #1 since Apr 16, 2026, 6 models in a row. We rank by the Artificial Analysis Intelligence Index. Other leaderboards, such as LMArena's human votes, crown different models.

  • OpenAI65%2.5 years at #1 across 12 models. Longest: GPT-4, 8 months.
  • Anthropic33%15 months at #1 across 11 models. Longest: Claude 3 Opus, 6 months.
  • xAI2%4 weeks at #1 across 1 model. Longest: Grok 4, 4 weeks.
  • Google<1%6 days at #1 across 1 model. Longest: Gemini 3 Pro Preview, 6 days.

The wider field

139 models were out on Mar 31, 2025

Best open model
DeepSeek R1 (Jan '25)

DeepSeek, scoring 11.4. #10 overall, 6.3 points behind.

Best under $1 per million
Grok 3 mini Reasoning

xAI, scoring 14.6. #4 overall at $0.35, 15× cheaper than #1.

New in the 30 days before · 17
  • GPT-4o (March 2025, chatgpt-4o-latest)9
  • Gemini 2.5 Pro Preview (Mar' 25)15
  • DeepSeek V3 03249.7
  • o1-pro12.4
  • Llama 3.3 Nemotron Super 49B v18.9
  • Mistral Small 3.17.1
and 11 more

Browse

The top AI model, month by month

2026
2025
2024
2023
2022

Try today's top models on your own prompt

Scores tell you who leads on average. Send one prompt to up to six models and see which answer you'd actually use, with what each one cost. For the long view of prices, read Six months later, the best AI costs 20× less.

Compare models

Scores and release dates from Artificial Analysis (Intelligence Index, best published effort of each model), data through October 18, 2026. Prices are list prices on the day, including later cuts we know of. Models without a public price are ranked but not priced.