Claude Opus 4.8 vs DeepSeek-V4.1-Flash
Wins 5 of 8 areas
Reasoning · Facts · Images and charts · Math · Following instructions
Wins 3 of 8 areas
Coding · Agents · Long documents
Claude Opus 4.8 is the stronger all-rounder.DeepSeek-V4.1-Flash is cheaper and better at agents.
Scores updated · 38 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.
- FactsGetting facts right instead of making them up30Claude Opus 4.83 of 3 tests
- MathCompetition and research-level math30Claude Opus 4.83 of 3 tests
- ReasoningHard problems that need careful thinking42Claude Opus 4.84 of 7 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video20Claude Opus 4.82 of 2 tests
- Following instructionsDoing exactly what it is asked10Claude Opus 4.81 of 1 test
- AgentsCarrying out multi-step tasks on its own02DeepSeek-V4.1-Flash2 of 2 tests
- CodingWriting and fixing software23DeepSeek-V4.1-Flash3 of 5 tests
- Long documentsFinding answers in very long texts01DeepSeek-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 4.8 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+57.2points ahead
- Reasoning about charts from research papersCharXiv (RQ)+20.9points ahead
- Short factual questions, answered correctlySimpleQA Verified+10.1points ahead
Where DeepSeek-V4.1-Flash pulls ahead
- Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+26.8points ahead
- Real work tasks from 44 professionsGDPVal+7.3points ahead
- Reasons across sets of long documentsAA-LCR+6.3points ahead
Every test, side by side
All 38 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-V4.1-Flash
- LiveBench · Agentic CodingDeepSeek-V4.1-Flash by 26.850.577.3+26.8
- LMArena · WebDevDeepSeek-V4.1-Flash by 63 rating points15561619+63 rating
- Terminal-Bench 2.1DeepSeek-V4.1-Flash by 684.690.6+6
- SciCodeClaude Opus 4.8 by 2.554.451.9+2.5
- LiveBench · CodingClaude Opus 4.8 by 1.881.880+1.8
AgentsDeepSeek-V4.1-Flash
- GDPValDeepSeek-V4.1-Flash by 7.347.755+7.3
- Terminal-Bench 4.0DeepSeek-V4.1-Flash by 5.121.726.8+5.1
ReasoningClaude Opus 4.8
- Humanity's Last ExamClaude Opus 4.8 by 9.548.739.2+9.5
- CritPtClaude Opus 4.8 by 6.620.914.3+6.6
- ARC-AGI-2DeepSeek-V4.1-Flash by 672.178.1+6
- LiveBench · ReasoningClaude Opus 4.8 by 2.589.286.7+2.5
- SimpleBenchDeepSeek-V4.1-Flash by 1.964.866.7+1.9
- GPQA DiamondClaude Opus 4.8 by 1.19290.9+1.1
- ARC-AGI-3tie1.50.9tie
FactsClaude Opus 4.8
- AA-Omniscience · Non-hallucinationClaude Opus 4.8 by 57.260.73.5+57.2
- SimpleQA VerifiedClaude Opus 4.8 by 10.15342.9+10.1
- AA-Omniscience · AccuracyClaude Opus 4.8 by 2.448.846.4+2.4
Images and chartsClaude Opus 4.8
- CharXiv (RQ)Claude Opus 4.8 by 20.989.969+20.9
- LMArena · VisionClaude Opus 4.8 by 16 rating points12941278+16 rating
MathClaude Opus 4.8
- FrontierMath Tier 4Claude Opus 4.8 by 9.856.146.3+9.8
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.8 by 8.88071.2+8.8
- LiveBench · MathematicsClaude Opus 4.8 by 194.393.3+1
Long documentsDeepSeek-V4.1-Flash
- AA-LCRDeepSeek-V4.1-Flash by 6.377.784+6.3
Following instructionsClaude Opus 4.8
- LiveBench · Instruction FollowingClaude Opus 4.8 by 27270+2
Other results14 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 4.8 by 51.671.920.3+51.6
- AA-OmniscienceClaude Opus 4.8 by 34.128.8-5.3+34.1
- DeepSWE 1.1DeepSeek-V4.1-Flash by 15.25974.2+15.2
- livebench_data_analysisDeepSeek-V4.1-Flash by 13.36679.3+13.3
- CyberGymDeepSeek-V4.1-Flash by 9.378.888.1+9.3
- Terminal-bench 3.0DeepSeek-V4.1-Flash by 8.921.130+8.9
- Terminal-Bench 4.0DeepSeek-V4.1-Flash by 7.623.631.2+7.6
- Agents' Last ExamDeepSeek-V4.1-Flash by 6.125.731.8+6.1
- HLE (with tools)DeepSeek-V4.1-Flash by 657.963.9+6
- NL2RepoClaude Opus 4.8 by 5.769.764+5.7
- AA IntelligenceClaude Opus 4.8 by 2.341.839.5+2.3
- livebench_languageDeepSeek-V4.1-Flash by 1.579.781.2+1.5
- OSWorld 2.0Claude Opus 4.8 by 1.455.754.3+1.4
- ARC-AGI-1DeepSeek-V4.1-Flash by 192.593.5+1
Questions people ask
Which is better, Claude Opus 4.8 or DeepSeek-V4.1-Flash?
Claude Opus 4.8 wins five of the eight areas we test: reasoning, facts, images and charts, math and following instructions. DeepSeek-V4.1-Flash wins coding, agents and long documents, and costs 95% less.
Which is better for coding?
DeepSeek-V4.1-Flash. It wins 3 of the 5 coding tests both models report; Claude Opus 4.8 wins 2.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; DeepSeek-V4.1-Flash costs $0.30 and $1.20. That makes DeepSeek-V4.1-Flash about 95% cheaper for the same work.
How do you compare the two?
We use the 38 benchmark tests both models have published scores on. The verdict counts the 24 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 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.