AI Benchmark Leaderboard

Compare AI model scores, coding ability and performance in one place.

Top models

The 8 highest-scoring AI models on the intelligence index, one per model family. OVR is that index, rounded; speed is output tokens per second and prices are US dollars per million tokens.

Ranked by Intelligence IndexCompare in the Arena

  1. 65OVRRank #1

    Claude Fable 5.2 Max, Default Fallback

    Anthropic

    Speed
    94tok/s
    First token
    2.1 min
    Input
    $8.00/1M
    Output
    $40.00/1M
  2. 60OVRRank #4

    Gemini 4 Carbon High

    Google

    Speed
    91tok/s
    First token
    117s
    Input
    $2.00/1M
    Output
    $10.00/1M
  3. 59OVRRank #5

    Claude Opus 5.5 Max, Default Fallback

    Anthropic

    Speed
    97tok/s
    First token
    11.7 min
    Input
    $4.00/1M
    Output
    $20.00/1M
  4. 57OVRRank #8

    GPT-6 Astra Max

    OpenAI

    Speed
    63tok/s
    First token
    4.9 min
    Input
    $10.00/1M
    Output
    $50.00/1M
  5. 55OVRRank #12

    GPT-6.1 Sol Max

    OpenAI

    Speed
    66tok/s
    First token
    4.6 min
    Input
    $2.00/1M
    Output
    $10.00/1M
  6. 55OVRRank #13

    Claude Sonnet 5.5 Max, Default Fallback

    Anthropic

    Speed
    133tok/s
    First token
    7.9 min
    Input
    $2.00/1M
    Output
    $10.00/1M
  7. 54OVRRank #15

    Gemini 4 Hydrogen Xhigh

    Google

    Speed
    239tok/s
    First token
    74s
    Input
    $1.50/1M
    Output
    $7.50/1M
  8. 51OVRRank #20

    GPT-6.1 Terra Max

    OpenAI

    Speed
    93tok/s
    First token
    5.5 min
    Input
    $1.00/1M
    Output
    $5.00/1M

# Model Org Intelligence ↓ CodingMathMMLU ProGPQALiveCodeBenchAIME 2025MATH 500SciCodeIFBenchHLE
1 Claude Fable 5.2 (Max, Default Fallback) Anthropic 64.6 ——————————
2 Claude Fable 5.2 (Xhigh, Default Fallback) Anthropic 62.2 ——————————
3 Claude Fable 5.2 (High, Default Fallback) Anthropic 61.8 ——————————
4 Gemini 4 Carbon (High) Google 60.0 ——————————
5 Claude Opus 5.5 (Max, Default Fallback) Anthropic 59.1 ——————————
6 Gemini 4 Carbon (Medium) Google 58.3 ——————————
7 Claude Opus 5.5 (Xhigh, Default Fallback) Anthropic 57.4 ——————————
8 GPT-6 Astra (Max) OpenAI 56.9 ——————————
9 Claude Fable 5.2 (Medium, Default Fallback) Anthropic 56.4 ——————————
10 GPT-6 Astra (Xhigh) OpenAI 55.3 ——————————
11 GPT-6 Astra (High) OpenAI 54.8 ——————————
12 GPT-6.1 Sol (Max) OpenAI 54.6 ——————————
13 Claude Sonnet 5.5 (Max, Default Fallback) Anthropic 54.5 ——————————
14 Gemini 4 Carbon (Low) Google 54.2 ——————————
15 Claude Fable 5.2 (Low, Default Fallback) Anthropic 53.9 ——————————
16 Gemini 4 Hydrogen (Xhigh) Google 53.9 ——————————
17 GPT-6.1 Sol (Xhigh) OpenAI 52.5 ——————————
18 Claude Opus 5.5 (High, Default Fallback) Anthropic 51.8 ——————————
19 GPT-6 Astra (Medium) OpenAI 51.6 ——————————
20 GPT-6.1 Terra (Max) OpenAI 51.3 ——————————
21 Gemini 4 Hydrogen (High) Google 50.8 ——————————
22 Gemini 4 Argon (High) Google 50.4 ——————————
23 GPT-6.1 Sol (High) OpenAI 50.3 ——————————
24 Muse Wonder Full (Max) Meta 50.1 ——————————
25 Claude Fable 5.1 (Xhigh, Default Fallback) Anthropic 50.0 ——————————

AI Benchmarks FAQ

Which AI model is best for my task?

Compare the scores most relevant to your task, then consider speed, pricing and practical requirements. A high score in one benchmark does not make a model the best choice for every use.

What does a missing benchmark score mean?

A dash means that a score is not available for that model in this table. It should not be treated as a score of zero.

How are the top models ranked?

By the intelligence index, highest first. The Top models cards show the eight highest-scoring current models, one per model family, and models with the same score share a place.

Where do the scores come from?

From independent third-party model evaluations, which Franklin AI syncs automatically each day. Each model page links to About these benchmarks, which explains them in more detail.

What does a model page show?

Its scores, its price per million input and output tokens, its measured speed and comparisons with other models, where that data is available.