Share promptsSign inCreate team library
The Chat Judge verdicts

Which AI is best at what?

Public benchmark results for 19 tests across 9 fields, charted side by side. Every score links to the leaderboard it came from.

Showing Google's best-ranked model on each benchmark. Try prompts for Gemini →

1 benchmarkUpdated 19 Oct 2026Overall workOne number for agentic work across finance, coding, law and tax.
  • Vals Index Gemini 4 Argon #3 · 69.24%
See charts 3 benchmarksUpdated 19 Oct 2026LegalDrafting, extraction, legal reasoning and US legal research.
  • Legal Benchmarks leaderboard Gemini 3.8 Flash #14 · 68.2%
  • LegalBench Gemini 4 Argon #3 · 88.30%
  • Legal Research Bench Gemini 4 Argon #4 · 54.81%
See charts 3 benchmarksUpdated 21 Oct 2026Health & medicalPhysician-graded health conversations, medical coding and exam questions.
  • HealthBench (independent run) Gemini 4 Argon (high) #13 · 49.1%
  • MedCode Gemini 3.1 Pro Preview (02/26) #2 · 59.06%
  • MedQA Gemini 3.1 Pro Preview (02/26) #3 · 96.37%
See charts 2 benchmarksUpdated 19 Oct 2026FinanceAnalyst research on SEC filings and long credit agreements.
  • Finance Agent Gemini 4 Argon #2 · 67.66%
  • CorpFin v2 Gemini 3 Flash (12/25) #20 · 66.43%
See charts 2 benchmarksUpdated 1 Sep 2026CodingFixing real GitHub issues and solving fresh programming problems.
  • SWE-bench Verified Gemini 3.7 Flash #23 · 80.8%
  • LiveCodeBench Gemini 3.8 Flash #3 · 89.48%
See charts 2 benchmarksUpdated 19 Oct 2026Agents & tool useMulti-step tasks an agent completes alone in a terminal.
  • Terminal-Bench 4.0 Gemini 4 Argon #6 · 57.58%
  • Terminal-Bench 2.1 Gemini 3.8 Flash #8 · 81.27%
See charts 4 benchmarksUpdated 21 Oct 2026Reasoning & knowledgeExpert exams, graduate science, contest maths and broad knowledge.
  • Humanity's Last Exam gemini-3.1-pro-preview (thinking high) #3 · 46.44%
  • GPQA Diamond Gemini 3.1 Pro Preview (02/26) #1 · 95.45%
  • AIME Gemini 3.1 Pro Preview (02/26) #1 · 98.13%
  • MMLU Pro Gemini 3.1 Pro Preview (02/26) #4 · 90.99%
See charts 1 benchmarkUpdated 20 Oct 2026Chat preferenceWhich answers real people prefer in blind side-by-side votes.
  • LMArena Text Arena Gemini 4 Argon (High) #1 · 1524
See charts 1 benchmarkUpdated 9 Jan 2026MultilingualReasoning in languages other than English.
  • MGSM Gemini 3 Pro (11/25) #7 · 93.93%
See charts

ChatJudge does not run these tests. Scores are copied as published by each source, with a link and the date we read them, and only change when we refresh from the source. Different leaderboards use different settings, so compare models within one chart, not across charts. Spotted a mistake? Tell us.