Claude Sonnet 4.6 vs Gemini 3.1 Pro
Wins 0 of 8 areas
—
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Following instructions
Gemini 3.1 Pro is the stronger all-rounder.
Scores updated · 70 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.
- ReasoningHard problems that need careful thinking04Gemini 3.1 Pro4 of 5 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video03Gemini 3.1 Pro3 of 3 tests
- AgentsCarrying out multi-step tasks on its own35Gemini 3.1 Pro5 of 8 tests
- MathCompetition and research-level math02Gemini 3.1 Pro2 of 2 tests
- Following instructionsDoing exactly what it is asked02Gemini 3.1 Pro2 of 2 tests
- CodingWriting and fixing software34Gemini 3.1 Pro4 of 8 tests · 1 tie
- FactsGetting facts right instead of making them up12Gemini 3.1 Pro2 of 4 tests · 1 tie
- Long documentsFinding answers in very long texts11Even1 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 22% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Gemini 3.1 Pro pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+40.7points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+20.5points ahead
- Proof problems from the US Math OlympiadUSAMO 2026+19.4points ahead
Where Claude Sonnet 4.6 pulls ahead
- Finds one of eight look-alike replies in a long chatMRCR v2 8 needle 128k (average)+58.6points ahead
- Real work tasks from 44 professionsGDPVal+22points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+13points ahead
Every test, side by side
All 70 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemini 3.1 Pro
- SciCodeGemini 3.1 Pro by 8.650.158.7+8.6
- LMArena · WebDevClaude Sonnet 4.6 by 75 rating points15221447+75 rating
- SWE-bench ProClaude Sonnet 4.6 by 3.958.154.2+3.9
- LiveBench · CodingClaude Sonnet 4.6 by 2.879.376.5+2.8
- Terminal-Bench 2.1Gemini 3.1 Pro by 2.671.273.8+2.6
- LiveBench · Agentic CodingGemini 3.1 Pro by 1.542.644.1+1.5
- SWE-bench VerifiedGemini 3.1 Pro by 179.680.6+1
- Terminal-Bench Hardtie5353.8tie
AgentsGemini 3.1 Pro
- GDPValClaude Sonnet 4.6 by 2236.714.7+22
- τ-Bench V3 · BankingClaude Sonnet 4.6 by 1334.421.4+13
- BrowseCompGemini 3.1 Pro by 11.274.785.9+11.2
- AA IT-Bench SREClaude Sonnet 4.6 by 9.539.830.3+9.5
- MCP AtlasGemini 3.1 Pro by 7.961.369.2+7.9
- AA ApexAgentsGemini 3.1 Pro by 42832+4
- OSWorld-VerifiedGemini 3.1 Pro by 3.772.576.2+3.7
- Terminal-Bench 4.0Gemini 3.1 Pro by 134+1
ReasoningGemini 3.1 Pro
- ARC-AGI-2Gemini 3.1 Pro by 18.858.377.1+18.8
- CritPtGemini 3.1 Pro by 14.63.117.7+14.6
- Humanity's Last ExamGemini 3.1 Pro by 13.433.647+13.4
- GPQA DiamondGemini 3.1 Pro by 6.687.594.1+6.6
- LiveBench · Reasoningtie84.884tie
FactsGemini 3.1 Pro
- SimpleQA VerifiedGemini 3.1 Pro by 40.732.873.5+40.7
- AA-Omniscience · AccuracyGemini 3.1 Pro by 1440.954.9+14
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.6 by 2.551.649.1+2.5
- Vectara HHEM hallucination ratelower is bettertie10.610.4tie
Images and chartsGemini 3.1 Pro
- CharXiv (RQ)Gemini 3.1 Pro by 11.771.683.3+11.7
- MMMU-ProGemini 3.1 Pro by 9.173.382.4+9.1
- LMArena · VisionGemini 3.1 Pro by 13 rating points12831296+13 rating
MathGemini 3.1 Pro
- USAMO 2026Gemini 3.1 Pro by 19.45574.4+19.4
- LiveBench · MathematicsGemini 3.1 Pro by 48791+4
Long documentsEven
- MRCR v2 8 needle 128k (average)Claude Sonnet 4.6 by 58.684.926.3+58.6
- AA-LCRGemini 3.1 Pro by 28082+2
Following instructionsGemini 3.1 Pro
- IFBenchGemini 3.1 Pro by 20.556.677.1+20.5
- LiveBench · Instruction FollowingGemini 3.1 Pro by 15.963.279.1+15.9
Other results36 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.
- GDP (Surge AI)Claude Sonnet 4.6 by 61.978.616.7+61.9
- GDPval-AA v2Claude Sonnet 4.6 by 433 rating points1395962+433 rating
- GDPval-AA (Elo)Claude Sonnet 4.6 by 359 rating points16761317+359 rating
- AA Agentic IndexClaude Sonnet 4.6 by 22.833.110.3+22.8
- τ²-Bench Telecom (AA run)Gemini 3.1 Pro by 19.975.795.6+19.9
- Blueprint-Bench 2Gemini 3.1 Pro by 19.86.726.5+19.8
- AA-OmniscienceGemini 3.1 Pro by 19.712.231.9+19.7
- OfficeQA ProGemini 3.1 Pro by 19.153.472.5+19.1
- DeepSWE 1.1Claude Sonnet 4.6 by 183012+18
- ARC-AGI-1Gemini 3.1 Pro by 128698+12
- Humanity’s Last ExamGemini 3.1 Pro by 11.233.244.4+11.2
- AgentWorldBench - MCPClaude Sonnet 4.6 by 10.97059.1+10.9
- Terminal-Bench 2.0Gemini 3.1 Pro by 9.459.168.5+9.4
- livebench_languageGemini 3.1 Pro by 9.376.185.4+9.3
- frontiermath_tier_4_v1Gemini 3.1 Pro by 8.48.316.7+8.4
- Finance Agent v2Claude Sonnet 4.6 by 85143+8
- ToolathlonGemini 3.1 Pro by 7.84148.8+7.8
- Artificial Analysis Coding IndexGemini 3.1 Pro by 5.86368.8+5.8
- AgentWorldBench - SWEClaude Sonnet 4.6 by 5.464.559.1+5.4
- Legal Agent BenchmarkClaude Sonnet 4.6 by 5.45.40+5.4
- HLE (with tools)Gemini 3.1 Pro by 4.646.851.4+4.6
- AgentWorldBench - TerminalClaude Sonnet 4.6 by 4.55752.5+4.5
- LiveBenchGemini 3.1 Pro by 4.475.579.9+4.4
- AgentWorldBench - OSGemini 3.1 Pro by 3.763.266.9+3.7
- Finance AgentClaude Sonnet 4.6 by 3.663.359.7+3.6
- AgentWorldBench - AndroidGemini 3.1 Pro by 3.45861.4+3.4
- MMMLUGemini 3.1 Pro by 3.389.392.6+3.3
- CharXiv Reasoning (With tools)Claude Sonnet 4.6 by 2.185.383.2+2.1
- AgentWorldBench - WebGemini 3.1 Pro by 250.852.8+2
- AgentWorldBench - OverallClaude Sonnet 4.6 by 1.45654.6+1.4
- AgentWorldBench - SearchGemini 3.1 Pro by 1.428.830.2+1.4
- vectara_avg_summary_lengthtie114.7107.7tie
- livebench_data_analysistie7878.5tie
- vectara_answer_ratetie99.999.4tie
- AA Intelligencetie30.129.7tie
- vectara_factual_consistencytie89.489.6tie
Questions people ask
Which is better, Claude Sonnet 4.6 or Gemini 3.1 Pro?
Gemini 3.1 Pro wins seven of the eight areas we test: coding, agents, reasoning, facts, images and charts, math and following instructions. Claude Sonnet 4.6 wins none. They are level on long documents.
Which is better for coding?
Gemini 3.1 Pro. It wins 4 of the 8 coding tests both models report; Claude Sonnet 4.6 wins 3, and 1 is a tie.
Which is cheaper?
Claude Sonnet 4.6 costs $3.00 per million input tokens and $15.00 per million output tokens; Gemini 3.1 Pro costs $2.00 and $12.00. That makes Gemini 3.1 Pro about 22% cheaper for the same work.
How do you compare the two?
We use the 70 benchmark tests both models have published scores on. The verdict counts the 34 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 36 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.