DeepSeek-V4.1-Flash vs Claude Opus 5
Wins 2 of 8 areas
Long documents · Following instructions
Wins 6 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math
Claude Opus 5 is the stronger all-rounder.DeepSeek-V4.1-Flash is cheaper and better at long documents.
Scores updated · 42 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 thinking07Claude Opus 57 of 7 tests
- FactsGetting facts right instead of making them up03Claude Opus 53 of 3 tests
- Images and chartsUnderstanding pictures, charts and video03Claude Opus 53 of 3 tests
- MathCompetition and research-level math03Claude Opus 53 of 3 tests
- AgentsCarrying out multi-step tasks on its own02Claude Opus 52 of 2 tests
- CodingWriting and fixing software23Claude Opus 53 of 5 tests
- Long documentsFinding answers in very long texts10DeepSeek-V4.1-Flash1 of 1 test
- Following instructionsDoing exactly what it is asked10DeepSeek-V4.1-Flash1 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.
DeepSeek-V4.1-Flash costs 95% 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 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+35.7points ahead
- Solves new puzzle games it has never seenARC-AGI-3+29.3points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+26.9points ahead
Where DeepSeek-V4.1-Flash pulls ahead
- Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+12.1points ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+6.2points ahead
- Reasons across sets of long documentsAA-LCR+4.7points ahead
Every test, side by side
All 42 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 5
- LiveBench · Agentic CodingDeepSeek-V4.1-Flash by 12.177.365.2+12.1
- LMArena · WebDevClaude Opus 5 by 71 rating points16191690+71 rating
- SciCodeClaude Opus 5 by 4.551.956.4+4.5
- LiveBench · CodingClaude Opus 5 by 1.58081.5+1.5
- Terminal-Bench 2.1DeepSeek-V4.1-Flash by 1.590.689.1+1.5
AgentsClaude Opus 5
- Terminal-Bench 4.0Claude Opus 5 by 22.226.849+22.2
- GDPValClaude Opus 5 by 6.15561.1+6.1
ReasoningClaude Opus 5
- ARC-AGI-3Claude Opus 5 by 29.30.930.2+29.3
- Humanity's Last ExamClaude Opus 5 by 15.739.254.9+15.7
- CritPtClaude Opus 5 by 14.814.329.1+14.8
- SimpleBenchClaude Opus 5 by 13.966.780.6+13.9
- ARC-AGI-2Claude Opus 5 by 12.378.190.4+12.3
- LiveBench · ReasoningClaude Opus 5 by 4.586.791.2+4.5
- GPQA DiamondClaude Opus 5 by 2.390.993.2+2.3
FactsClaude Opus 5
- AA-Omniscience · Non-hallucinationClaude Opus 5 by 35.73.539.2+35.7
- SimpleQA VerifiedClaude Opus 5 by 1742.959.9+17
- AA-Omniscience · AccuracyClaude Opus 5 by 14.546.460.9+14.5
Images and chartsClaude Opus 5
- CharXiv (RQ)Claude Opus 5 by 14.76983.7+14.7
- MMMU-ProClaude Opus 5 by 7.77784.7+7.7
- LMArena · VisionClaude Opus 5 by 42 rating points12781320+42 rating
MathClaude Opus 5
- FrontierMath Tier 4Claude Opus 5 by 26.946.373.2+26.9
- FrontierMath Tiers 1-3 (v2)Claude Opus 5 by 14.471.285.6+14.4
- LiveBench · MathematicsClaude Opus 5 by 2.493.395.7+2.4
Long documentsDeepSeek-V4.1-Flash
- AA-LCRDeepSeek-V4.1-Flash by 4.78479.3+4.7
Following instructionsDeepSeek-V4.1-Flash
- LiveBench · Instruction FollowingDeepSeek-V4.1-Flash by 6.27063.8+6.2
Other results17 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.
- Program BenchClaude Opus 5 by 65.120.385.4+65.1
- AA-OmniscienceClaude Opus 5 by 42.4-5.337.1+42.4
- Terminal-Bench 4.0Claude Opus 5 by 20.631.251.8+20.6
- SWE Refactor BenchClaude Opus 5 by 19.211.831+19.2
- OSWorld 2.0Claude Opus 5 by 16.354.370.6+16.3
- AA IntelligenceClaude Opus 5 by 11.339.550.8+11.3
- LoCoMoClaude Opus 5 by 10.758.869.5+10.7
- PostTrain Bench v1.1Claude Opus 5 by 9.930.540.4+9.9
- livebench_languageClaude Opus 5 by 7.581.288.7+7.5
- DeepSWE 1.1DeepSeek-V4.1-Flash by 5.474.268.8+5.4
- Chartography w/ tools (Pass@1)Claude Opus 5 by 5.178.984+5.1
- livebench_data_analysisDeepSeek-V4.1-Flash by 4.779.374.5+4.7
- BabyVisionClaude Opus 5 by 4.589.694.1+4.5
- ARC-AGI-1Claude Opus 5 by 493.597.5+4
- Agents' Last ExamDeepSeek-V4.1-Flash by 3.231.828.6+3.2
- ZeroBench-main (w/tools)Claude Opus 5 by 34952+3
- HLE (with tools)tie63.963.6tie
Questions people ask
Which is better, DeepSeek-V4.1-Flash or Claude Opus 5?
Claude Opus 5 wins six of the eight areas we test: coding, agents, reasoning, facts, images and charts and math. DeepSeek-V4.1-Flash wins long documents and following instructions, and costs 95% less.
Which is better for coding?
Claude Opus 5. It wins 3 of the 5 coding tests both models report; DeepSeek-V4.1-Flash wins 2.
Which is cheaper?
DeepSeek-V4.1-Flash costs $0.30 per million input tokens and $1.20 per million output tokens; Claude Opus 5 costs $5.00 and $25.00. That makes DeepSeek-V4.1-Flash about 95% cheaper for the same work.
How do you compare the two?
We use the 42 benchmark tests both models have published scores on. The verdict counts the 25 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 17 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.