Claude Opus 4.8 vs Gemini 3.7 Flash
Wins 2 of 8 areas
Agents · Math
Wins 4 of 8 areas
Coding · Facts · Long documents · Following instructions
Gemini 3.7 Flash wins more areas, narrowly.Claude Opus 4.8 is better at agents.
Scores updated · 39 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 software14Gemini 3.7 Flash4 of 5 tests
- FactsGetting facts right instead of making them up12Gemini 3.7 Flash2 of 3 tests
- Long documentsFinding answers in very long texts01Gemini 3.7 Flash1 of 1 test
- Following instructionsDoing exactly what it is asked01Gemini 3.7 Flash1 of 1 test
- AgentsCarrying out multi-step tasks on its own30Claude Opus 4.83 of 3 tests
- MathCompetition and research-level math20Claude Opus 4.82 of 3 tests · 1 tie
- ReasoningHard problems that need careful thinking22Even2 each · 1 tie
- Images and chartsUnderstanding pictures, charts and video11Even1 each
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 3.7 Flash costs 85% 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 3.7 Flash pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+16.2points ahead
- Abstract visual puzzles that people can solveARC-AGI-2+12.5points ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+7.9points ahead
Where Claude Opus 4.8 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+25.2points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+19.5points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+8.4points ahead
Every test, side by side
All 39 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemini 3.7 Flash
- LiveBench · Agentic CodingGemini 3.7 Flash by 7.850.558.3+7.8
- LMArena · WebDevGemini 3.7 Flash by 36 rating points15561592+36 rating
- LiveBench · CodingClaude Opus 4.8 by 2.981.878.9+2.9
- SciCodeGemini 3.7 Flash by 2.854.457.2+2.8
- Terminal-Bench 2.1Gemini 3.7 Flash by 1.284.685.8+1.2
AgentsClaude Opus 4.8
- Terminal-Bench 4.0Claude Opus 4.8 by 8.121.713.6+8.1
- GDPValClaude Opus 4.8 by 3.247.844.6+3.2
- τ-Bench V3 · BankingClaude Opus 4.8 by 1.434.232.8+1.4
ReasoningEven
- ARC-AGI-2Gemini 3.7 Flash by 12.572.184.6+12.5
- CritPtClaude Opus 4.8 by 6.620.914.3+6.6
- GPQA DiamondGemini 3.7 Flash by 2.59294.5+2.5
- LiveBench · ReasoningClaude Opus 4.8 by 1.489.287.8+1.4
- Humanity's Last Examtie48.747.9tie
FactsGemini 3.7 Flash
- AA-Omniscience · Non-hallucinationClaude Opus 4.8 by 25.260.735.5+25.2
- SimpleQA VerifiedGemini 3.7 Flash by 16.25369.2+16.2
- AA-Omniscience · AccuracyGemini 3.7 Flash by 6.548.855.3+6.5
Images and chartsEven
- LMArena · VisionGemini 3.7 Flash by 20 rating points12951315+20 rating
- CharXiv (RQ)Claude Opus 4.8 by 1.289.988.7+1.2
MathClaude Opus 4.8
- FrontierMath Tier 4Claude Opus 4.8 by 19.556.136.6+19.5
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.8 by 8.48071.6+8.4
- LiveBench · Mathematicstie94.393.5tie
Following instructionsGemini 3.7 Flash
- LiveBench · Instruction FollowingGemini 3.7 Flash by 7.97279.9+7.9
Other results16 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 Opus 4.8 by 50.884.834+50.8
- Terminal-Bench 4.0Claude Opus 4.8 by 12.423.611.2+12.4
- GDPval-AA v2 EloClaude Opus 4.8 by 111 rating points15931482+111 rating
- OSWorld 2.0Claude Opus 4.8 by 7.855.747.9+7.8
- GDPval-AA v2Claude Opus 4.8 by 68 rating points15931525+68 rating
- DeepSWE 1.1Gemini 3.7 Flash by 6.35965.3+6.3
- AA Agentic IndexClaude Opus 4.8 by 6.242.636.4+6.2
- Terminal-bench 3.0Claude Opus 4.8 by 6.221.114.9+6.2
- livebench_languageGemini 3.7 Flash by 5.879.785.5+5.8
- ARC-AGI-1Gemini 3.7 Flash by 392.595.5+3
- AA IntelligenceClaude Opus 4.8 by 2.741.839.1+2.7
- AA-OmniscienceClaude Opus 4.8 by 2.328.826.5+2.3
- livebench_data_analysisGemini 3.7 Flash by 26668+2
- Artificial Analysis Coding IndexGemini 3.7 Flash by 1.874.376.1+1.8
- CharXiv Reasoning (With tools)Claude Opus 4.8 by 1.289.988.7+1.2
- Agents' Last Examtie25.726.3tie
Questions people ask
Which is better, Claude Opus 4.8 or Gemini 3.7 Flash?
Gemini 3.7 Flash wins four of the eight areas we test: coding, facts, long documents and following instructions. Claude Opus 4.8 wins agents and math. They are level on reasoning and images and charts.
Which is better for coding?
Gemini 3.7 Flash. It wins 4 of the 5 coding tests both models report; Claude Opus 4.8 wins 1.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Gemini 3.7 Flash costs $0.75 and $3.75. That makes Gemini 3.7 Flash about 85% cheaper for the same work.
How do you compare the two?
We use the 39 benchmark tests both models have published scores on. The verdict counts the 23 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 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.