in the loop

Gemini 3 Flash
#48 of 274 scored.

GoogleDec 17, 2025Closed

Capabilities Index

Range 150.3–153.3

151.8

Context

Tokens

—

Input

Per 1M tokens

—

Output

Per 1M tokens

—

Benchmarks

How it scores.

  • GPQA Diamond85.9%

    43rd of 186

  • SWE-bench Verified75.4%

    13th of 32

  • FrontierMath51.2%

    45th of 81

  • ARC-AGI-233.6%

    35th of 79

  • SimpleQA Verified66.8%

    10th of 77

  • Mock AIME95.5%

    32nd of 176

  • Terminal-Bench64.3%

    9th of 35

Everything else Epoch has for it

  • GeoBench88.0%
  • ARC-AGI84.7%
  • DTBench81.8%
  • WeirdML61.6%
  • VPCT58.9%
  • SimpleBench53.3%
  • LMCA50.7%
  • DeepResearch Bench49.8%
  • Balrog48.1%
  • Chess Puzzles36.9%
  • Mystery Game Puzzles18.5%
  • FrontierMath-Tier-4-v2-Private17.1%
  • ProofBench15.0%
  • GSO-Bench9.8%

On the index

Around it.

  1. #45Muse SparkMeta152.0
  2. #46Grok 4.20xAI · $1.25 in152.0
  3. #47GLM-5.3-FlashZ.ai · $0.15 in151.9
  4. #48Gemini 3 FlashGoogle151.8
  5. #49GLM-5.2Z.ai · $0.06 in151.8
  6. #50Kimi K2.6Moonshot · $0.47 in151.1
  7. #51GPT-5 ProOpenAI · $15 in150.3

Sources