DeepSeek-V4-Pro vs Claude Opus 5
Wins 0 of 7 areas
—
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Long documents · Following instructions
Claude Opus 5 is the stronger all-rounder.DeepSeek-V4-Pro is cheaper.
Scores updated · 39 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. 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 software08Claude Opus 58 of 8 tests
- AgentsCarrying out multi-step tasks on its own05Claude Opus 55 of 5 tests
- ReasoningHard problems that need careful thinking05Claude Opus 55 of 5 tests
- FactsGetting facts right instead of making them up03Claude Opus 53 of 3 tests
- MathCompetition and research-level math03Claude Opus 53 of 3 tests
- Long documentsFinding answers in very long texts01Claude Opus 51 of 1 test
- Following instructionsDoing exactly what it is asked01Claude Opus 51 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-Pro costs 96% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Claude Opus 5 pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+70.8points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+40.3points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+34.4points ahead
Where DeepSeek-V4-Pro pulls ahead
No clear win on a test scored out of 100.
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.
CodingClaude Opus 5
- Terminal-Bench 2.1Claude Opus 5 by 25.16489.1+25.1
- SWE-bench ProClaude Opus 5 by 23.855.479.2+23.8
- LMArena · WebDevClaude Opus 5 by 228 rating points14641692+228 rating
- LiveBench · Agentic CodingClaude Opus 5 by 22.642.665.2+22.6
- SWE-bench VerifiedClaude Opus 5 by 15.480.696+15.4
- SWE-bench MultilingualClaude Opus 5 by 13.376.289.5+13.3
- LiveBench · CodingClaude Opus 5 by 11.57081.5+11.5
- SciCodeClaude Opus 5 by 5.650.856.4+5.6
AgentsClaude Opus 5
- Terminal-Bench 4.0Claude Opus 5 by 34.414.649+34.4
- GDPValClaude Opus 5 by 28.332.961.2+28.3
- MCP AtlasClaude Opus 5 by 12.273.685.8+12.2
- τ-Bench V3 · BankingClaude Opus 5 by 1230.142.1+12
- BrowseCompClaude Opus 5 by 7.483.490.8+7.4
ReasoningClaude Opus 5
- SimpleBenchClaude Opus 5 by 29.750.980.6+29.7
- Humanity's Last ExamClaude Opus 5 by 17.437.554.9+17.4
- CritPtClaude Opus 5 by 16.212.929.1+16.2
- LiveBench · ReasoningClaude Opus 5 by 8.582.791.2+8.5
- GPQA DiamondClaude Opus 5 by 4.488.893.2+4.4
FactsClaude Opus 5
- AA-Omniscience · Non-hallucinationClaude Opus 5 by 33.35.939.2+33.3
- AA-Omniscience · AccuracyClaude Opus 5 by 17.94360.9+17.9
- SimpleQA VerifiedClaude Opus 5 by 13.746.259.9+13.7
MathClaude Opus 5
- FrontierMath Tier 4Claude Opus 5 by 70.82.473.2+70.8
- FrontierMath Tiers 1-3 (v2)Claude Opus 5 by 40.345.385.6+40.3
- LiveBench · MathematicsClaude Opus 5 by 590.795.7+5
Following instructionsClaude Opus 5
- LiveBench · Instruction FollowingClaude Opus 5 by 1.462.463.8+1.4
Other results13 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.
- GDPval-AA v2Claude Opus 5 by 518 rating points13061824+518 rating
- AA-OmniscienceClaude Opus 5 by 47.8-10.737.1+47.8
- Terminal-Bench 4.0Claude Opus 5 by 39.412.451.8+39.4
- Program BenchClaude Opus 5 by 37.647.885.4+37.6
- AA Agentic IndexClaude Opus 5 by 28.527.756.2+28.5
- Toolathlon VerifiedClaude Opus 5 by 24.755.980.6+24.7
- AA IntelligenceClaude Opus 5 by 20.430.450.8+20.4
- Artificial Analysis Coding IndexClaude Opus 5 by 18.659.478+18.6
- GDPval-AA (Elo)Claude Opus 5 by 154 rating points15541708+154 rating
- livebench_languageClaude Opus 5 by 10.678.188.7+10.6
- Agents' Last ExamClaude Opus 5 by 5.925.731.6+5.9
- HLE (with tools)Claude Opus 5 by 3.66063.6+3.6
- livebench_data_analysistie74.574.5tie
Questions people ask
Which is better, DeepSeek-V4-Pro or Claude Opus 5?
Claude Opus 5 wins all seven areas where both have results: coding, agents, reasoning, facts, math, long documents and following instructions. DeepSeek-V4-Pro wins none, but costs 96% less.
Which is better for coding?
Claude Opus 5. It wins 8 of the 8 coding tests both models report; DeepSeek-V4-Pro wins none.
Which is cheaper?
DeepSeek-V4-Pro costs $0.43 per million input tokens and $0.87 per million output tokens; Claude Opus 5 costs $5.00 and $25.00. That makes DeepSeek-V4-Pro about 96% 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 26 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 13 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.