Claude Opus 5.5 vs Gemini 3.1 Pro
Wins 6 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents
Wins 2 of 8 areas
Facts · Following instructions
Claude Opus 5.5 is the stronger all-rounder.Gemini 3.1 Pro is cheaper and better at facts.
Both rank among the ten best models we track in facts.
Scores updated · 49 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software60Claude Opus 5.56 of 6 tests
- ReasoningHard problems that need careful thinking50Claude Opus 5.55 of 5 tests
- AgentsCarrying out multi-step tasks on its own40Claude Opus 5.54 of 4 tests
- MathCompetition and research-level math30Claude Opus 5.53 of 3 tests
- Images and chartsUnderstanding pictures, charts and video20Claude Opus 5.52 of 2 tests
- Long documentsFinding answers in very long texts10Claude Opus 5.51 of 1 test
- FactsGetting facts right instead of making them up12Gemini 3.1 Pro2 of 4 tests · 1 tie
- Following instructionsDoing exactly what it is asked01Gemini 3.1 Pro1 of 1 test
Long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
Gemini 3.1 Pro costs 42% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Claude Opus 5.5 pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+68.2points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+55.6points ahead
- Real work tasks from 44 professionsGDPVal+53.7points ahead
Where Gemini 3.1 Pro pulls ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+13.4points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+7.7points ahead
- Short factual questions, answered correctlySimpleQA Verified+1.3points ahead
Every test, side by side
All 49 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 5.5
- LMArena · WebDevClaude Opus 5.5 by 369 rating points18151446+369 rating
- SWE-bench ProClaude Opus 5.5 by 35.789.954.2+35.7
- LiveBench · Agentic CodingClaude Opus 5.5 by 27.671.744.1+27.6
- SWE-bench MultilingualClaude Opus 5.5 by 1793.976.9+17
- LiveBench · CodingClaude Opus 5.5 by 12.889.376.5+12.8
- SciCodeClaude Opus 5.5 by 8.266.958.7+8.2
AgentsClaude Opus 5.5
- Terminal-Bench 4.0Claude Opus 5.5 by 55.659.64+55.6
- GDPValClaude Opus 5.5 by 53.768.314.6+53.7
- AA IT-Bench SREClaude Opus 5.5 by 7.938.230.3+7.9
- BrowseCompClaude Opus 5.5 by 4.990.885.9+4.9
ReasoningClaude Opus 5.5
- ARC-AGI-2Claude Opus 5.5 by 14.691.777.1+14.6
- Humanity's Last ExamClaude Opus 5.5 by 14.461.447+14.4
- CritPtClaude Opus 5.5 by 1431.717.7+14
- SimpleBenchClaude Opus 5.5 by 8.888.479.6+8.8
- LiveBench · ReasoningClaude Opus 5.5 by 8.292.284+8.2
FactsGemini 3.1 Pro
- AA-Omniscience · AccuracyClaude Opus 5.5 by 11.366.254.9+11.3
- AA-Omniscience · Non-hallucinationGemini 3.1 Pro by 7.741.449.1+7.7
- SimpleQA VerifiedGemini 3.1 Pro by 1.372.273.5+1.3
- Vectara HHEM hallucination ratelower is bettertie9.910.4tie
Images and chartsClaude Opus 5.5
- MMMU-ProClaude Opus 5.5 by 5.387.782.4+5.3
- LMArena · VisionClaude Opus 5.5 by 28 rating points13241296+28 rating
MathClaude Opus 5.5
- FrontierMath Tier 4Claude Opus 5.5 by 68.29526.8+68.2
- FrontierMath Tiers 1-3 (v2)Claude Opus 5.5 by 31.591.259.6+31.5
- LiveBench · MathematicsClaude Opus 5.5 by 6.197.191+6.1
Following instructionsGemini 3.1 Pro
- LiveBench · Instruction FollowingGemini 3.1 Pro by 13.465.779.1+13.4
Other results23 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- DeepSWE 1.1Claude Opus 5.5 by 62.274.212+62.2
- Program BenchClaude Opus 5.5 by 51.791.239.5+51.7
- GDPval-AA (Elo)Claude Opus 5.5 by 503 rating points18201317+503 rating
- HLE-DiamondClaude Opus 5.5 by 36.567.931.4+36.5
- baby_vision_with_pythonClaude Opus 5.5 by 29.497.768.3+29.4
- AA IntelligenceClaude Opus 5.5 by 27.957.629.7+27.9
- PostTrainBenchClaude Opus 5.5 by 27.749.321.6+27.7
- ArXivMath without toolsClaude Opus 5.5 by 25.291.266+25.2
- HiL-Bench (Tools-allowed)Claude Opus 5.5 by 23.75935.3+23.7
- LVBenchClaude Opus 5.5 by 17.583.766.2+17.5
- HLE (with tools)Claude Opus 5.5 by 16.367.751.4+16.3
- AA-OmniscienceClaude Opus 5.5 by 14.546.431.9+14.5
- vectara_avg_summary_lengthClaude Opus 5.5 by 84.4 rating points192.1107.7+84.4 rating
- GDP (Surge AI)Claude Opus 5.5 by 8.426.217.8+8.4
- Legal Agent BenchmarkClaude Opus 5.5 by 8.38.30+8.3
- OfficeQA ProGemini 3.1 Pro by 4.867.772.5+4.8
- mmmu_pro_with_pythonClaude Opus 5.5 by 4.589.885.3+4.5
- charxiv_rq_with_pythonClaude Opus 5.5 by 3.993.889.9+3.9
- livebench_data_analysisClaude Opus 5.5 by 1.880.378.5+1.8
- livebench_languagetie86.385.4tie
- ARC-AGI-1tie97.598tie
- vectara_factual_consistencytie90.189.6tie
- vectara_answer_ratetie99.799.4tie
Questions people ask
Which is better, Claude Opus 5.5 or Gemini 3.1 Pro?
Claude Opus 5.5 wins six of the eight areas we test: coding, agents, reasoning, images and charts, math and long documents. Gemini 3.1 Pro wins facts and following instructions, and costs 42% less.
Which is better for coding?
Claude Opus 5.5. It wins 6 of the 6 coding tests both models report; Gemini 3.1 Pro wins none.
Which is cheaper?
Claude Opus 5.5 costs $4.00 per million input tokens and $20.00 per million output tokens; Gemini 3.1 Pro costs $2.00 and $12.00. That makes Gemini 3.1 Pro about 42% cheaper for the same work.
How do you compare the two?
We use the 49 benchmark tests both models have published scores on. The verdict counts the 26 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 23 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.