Claude Opus 4.7 vs Gemini 3.5 Flash
Wins 4 of 8 areas
Coding · Agents · Math · Long documents
Wins 2 of 8 areas
Facts · Following instructions
Claude Opus 4.7 wins more areas, narrowly.Gemini 3.5 Flash is cheaper and better at following instructions.
Scores updated · 55 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 4.76 of 7 tests · 1 tie
- MathCompetition and research-level math31Claude Opus 4.73 of 5 tests · 1 tie
- Long documentsFinding answers in very long texts20Claude Opus 4.72 of 2 tests
- AgentsCarrying out multi-step tasks on its own32Claude Opus 4.73 of 6 tests · 1 tie
- Following instructionsDoing exactly what it is asked02Gemini 3.5 Flash2 of 2 tests
- FactsGetting facts right instead of making them up12Gemini 3.5 Flash2 of 3 tests
- ReasoningHard problems that need careful thinking22Even2 each · 2 ties
- Images and chartsUnderstanding pictures, charts and video11Even1 each · 1 tie
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 3.5 Flash costs 65% 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 4.7 pulls ahead
- Finds one of eight look-alike replies in a long chatMRCR v2 8 needle 128k (average)+32.7points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+19.9points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+10.6points ahead
Where Gemini 3.5 Flash pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+17.7points ahead
- Common-sense trick questionsSimpleBench+15points ahead
- Short factual questions, answered correctlySimpleQA Verified+14.5points ahead
Every test, side by side
All 55 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4.7
- Terminal-Bench HardClaude Opus 4.7 by 10.651.540.9+10.6
- SWE-bench ProClaude Opus 4.7 by 9.264.355.1+9.2
- LMArena · WebDevClaude Opus 4.7 by 59 rating points15581499+59 rating
- Terminal-Bench 2.1Claude Opus 4.7 by 4.483.178.7+4.4
- LiveBench · CodingClaude Opus 4.7 by 3.982.178.2+3.9
- LiveBench · Agentic CodingClaude Opus 4.7 by 1.750.749+1.7
- SciCodetie54.553.9tie
AgentsClaude Opus 4.7
- AA ApexAgentsGemini 3.5 Flash by 13.233.947.1+13.2
- GDPValClaude Opus 4.7 by 7.642.835.2+7.6
- AA IT-Bench SREClaude Opus 4.7 by 6.446.740.3+6.4
- MCP AtlasGemini 3.5 Flash by 6.377.383.6+6.3
- τ-Bench V3 · BankingClaude Opus 4.7 by 2.434.632.2+2.4
- OSWorld-Verifiedtie7878.4tie
ReasoningEven
- SimpleBenchGemini 3.5 Flash by 1561.776.7+15
- LiveBench · ReasoningClaude Opus 4.7 by 5.287.282+5.2
- ARC-AGI-2Claude Opus 4.7 by 3.775.872.1+3.7
- CritPtGemini 3.5 Flash by 1.11213.1+1.1
- GPQA Diamondtie91.492.2tie
- Humanity's Last Examtie42.342.7tie
FactsGemini 3.5 Flash
- AA-Omniscience · Non-hallucinationClaude Opus 4.7 by 19.957.737.8+19.9
- SimpleQA VerifiedGemini 3.5 Flash by 14.551.766.2+14.5
- AA-Omniscience · AccuracyGemini 3.5 Flash by 2.548.951.4+2.5
Images and chartsEven
- CharXiv (RQ)Claude Opus 4.7 by 6.89184.2+6.8
- MMMU-ProGemini 3.5 Flash by 5.578.884.3+5.5
- LMArena · Visiontie13131310tie
MathClaude Opus 4.7
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.7 by 7.470.262.8+7.4
- FrontierMath Tier 4Claude Opus 4.7 by 4.931.726.8+4.9
- LiveBench · MathematicsClaude Opus 4.7 by 4.792.888.2+4.7
- HMMT Feb. 2026Gemini 3.5 Flash by 1.693.995.5+1.6
- AIME 2026tie95.895tie
Long documentsClaude Opus 4.7
- MRCR v2 8 needle 128k (average)Claude Opus 4.7 by 32.759.326.6+32.7
- AA-LCRClaude Opus 4.7 by 5.478.773.3+5.4
Following instructionsGemini 3.5 Flash
- IFBenchGemini 3.5 Flash by 17.758.676.3+17.7
- LiveBench · Instruction FollowingGemini 3.5 Flash by 8.966.775.6+8.9
Other results21 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.
- DeepSWEClaude Opus 4.7 by 175437+17
- HiL-Bench (Tools-allowed)Claude Opus 4.7 by 1441.727.7+14
- livebench_data_analysisClaude Opus 4.7 by 13.478.364.9+13.4
- AA Agentic IndexClaude Opus 4.7 by 12.239.527.3+12.2
- GDPval-AA (Elo)Claude Opus 4.7 by 97 rating points17531656+97 rating
- Blueprint-Bench 2Gemini 3.5 Flash by 9.124.533.6+9.1
- frontiermath_tier_4_v1Claude Opus 4.7 by 8.322.914.6+8.3
- AA IntelligenceClaude Opus 4.7 by 8.140.732.6+8.1
- Terminal-Bench 2.0Gemini 3.5 Flash by 6.869.476.2+6.8
- Humanity’s Last ExamClaude Opus 4.7 by 6.746.940.2+6.7
- livebench_languageGemini 3.5 Flash by 6.777.984.6+6.7
- τ²-Bench Telecom (AA run)Gemini 3.5 Flash by 6.788.695.3+6.7
- Finance AgentClaude Opus 4.7 by 6.564.457.9+6.5
- Finance Agent v2Gemini 3.5 Flash by 6.451.557.9+6.4
- Legal Agent BenchmarkClaude Opus 4.7 by 6.37.10.8+6.3
- AA-OmniscienceClaude Opus 4.7 by 6.127.321.2+6.1
- CharXiv Reasoning (With tools)Claude Opus 4.7 by 6.19184.9+6.1
- Artificial Analysis Coding IndexClaude Opus 4.7 by 3.573.670.1+3.5
- ToolathlonClaude Opus 4.7 by 2.859.356.5+2.8
- LiveBenchClaude Opus 4.7 by 1.976.975+1.9
- ARC-AGI-1tie9292.5tie
Questions people ask
Which is better, Claude Opus 4.7 or Gemini 3.5 Flash?
Claude Opus 4.7 wins four of the eight areas we test: coding, agents, math and long documents. Gemini 3.5 Flash wins facts and following instructions, and costs 65% less. They are level on reasoning and images and charts.
Which is better for coding?
Claude Opus 4.7. It wins 6 of the 7 coding tests both models report; Gemini 3.5 Flash wins none, and 1 is a tie.
Which is cheaper?
Claude Opus 4.7 costs $5.00 per million input tokens and $25.00 per million output tokens; Gemini 3.5 Flash costs $1.50 and $9.00. That makes Gemini 3.5 Flash about 65% cheaper for the same work.
How do you compare the two?
We use the 55 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 21 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.