in the loop

Gemini 2.5 Pro (May 2025)
#103 of 274 scored.

GoogleMay 6, 2025Closed

Capabilities Index

Range 139.0–145.1

142.5

Context

Tokens

—

Input

Per 1M tokens

—

Output

Per 1M tokens

—

Benchmarks

How it scores.

  • GPQA Diamond55.6%

    107th of 186

  • Humanity's Last Exam13.7%

    22nd of 40

Everything else Epoch has for it

  • MATH level 595.9%
  • GeoBench86.0%
  • Lech Mazur Writing80.9%
  • Aider polyglot76.9%
  • Fiction.LiveBench66.7%
  • The Agent Company30.3%
  • VPCT10.8%

On the index

Around it.

  1. #100Claude Opus 4Anthropic142.7
  2. #101GPT-5.5 InstantOpenAI142.5
  3. #102Qwen3.5-35B-A3BAlibaba · $0.08 in142.5
  4. #103Gemini 2.5 Pro (May 2025)Google142.5
  5. #104Claude Haiku 4.5Anthropic · $1.00 in142.4
  6. #105Qwen3-MaxAlibaba142.4
  7. #106o1OpenAI · $15 in141.9

Sources