Live ranking

The best AI model

Our aggregated ranking of frontier AI models, refreshed every couple of hours. Scored across the benchmarks that matter — Arena Elo, GPQA Diamond, SWE-bench Verified, LiveBench, Terminal-Bench and the Artificial Analysis Intelligence Index.

Current #1

Anthropic: Claude Fable 5

Anthropic

Top 10

  1. #01Anthropic: Claude Fable 5
    Elo 1572
  2. #02MoonshotAI: Kimi K3
    Elo 1558
  3. #03OpenAI: GPT-5.6 Sol
    Elo 1560
  4. #04xAI: Grok 4.5
    Elo 1564
  5. #05DeepSeek: DeepSeek V4 Pro
    Elo 1539
  6. #06Qwen: Qwen3.7 Max
    Elo 1526
  7. #07Anthropic: Claude Opus 4.8
    Elo 1548
  8. #08Anthropic: Claude Opus 4.8 (Fast)
    Elo 1548
  9. #09Z.ai: GLM 5.2
    Elo 1518
  10. #10NVIDIA: Nemotron 3 Ultra
    Elo 1509

Frequently asked

What is the best AI model in 2026?
The current state-of-the-art AI model on our aggregated leaderboard is shown at the top of this page. Rankings are pulled live from OpenRouter and scored across LMArena, Artificial Analysis, LiveBench, GPQA, SWE-bench and more.
Best AI model vs best LLM — is there a difference?
In practice, no — 'AI model' as searched almost always means large language model. If you need multimodal (image, audio, video generation), check the modality column on each model page.