BulletBench: Ladder Elo

Top 10 of 45 models · ladder Elo, higher is better · lines show the 95% range · ≈ cannot be told apart from the leader

  1. 1≈ Gemini 3.5 Flash-LiteGoogle 892 (95% range 641–1,081)
  2. 2≈ Gemini 3.1 Flash-Lite (minimal reasoning)Google 780 (95% range 589–940)
  3. 3≈ Gemini Decision 1Google 755 (95% range 535–964)
  4. 4≈ Gemini 3.5 Flash (minimal reasoning)Google 727 (95% range 567–890)
  5. 5≈ Clef OmniCloudflare 616 (95% range 466–761)
  6. 5≈ ReflexMorph 616 (95% range 470–757)
  7. 7≈ Elephant 1Baseten 614 (95% range 470–746)
  8. 7≈ Mercury DecideInception 614 (95% range 470–766)
  9. 7≈ Drex Max FinanceNace AI 614 (95% range 429–742)
  10. 7≈ Span-02Respan 614 (95% range 429–742)

Results to 12 Nov 2026 · release 3608133e8331Source: Spring Prompt · CC BY 4.0