Claude Opus 4.8 vs DeepSeek-V4-Pro
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Long documents
Wins 0 of 7 areas
—
Claude Opus 4.8 is the stronger all-rounder.DeepSeek-V4-Pro is cheaper.
Scores updated · 68 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 software90Claude Opus 4.89 of 9 tests
- AgentsCarrying out multi-step tasks on its own50Claude Opus 4.85 of 6 tests · 1 tie
- ReasoningHard problems that need careful thinking50Claude Opus 4.85 of 5 tests
- MathCompetition and research-level math51Claude Opus 4.85 of 7 tests · 1 tie
- FactsGetting facts right instead of making them up30Claude Opus 4.83 of 3 tests
- Long documentsFinding answers in very long texts10Claude Opus 4.81 of 1 test
- Following instructionsDoing exactly what it is asked11Even1 each
Long documents rests on a single test.
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 4.8 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+54.8points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+53.7points ahead
- Proof problems from the US Math OlympiadUSAMO 2026+36points ahead
Where DeepSeek-V4-Pro pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+14.3points ahead
- Olympiad math problems with short answersIMOAnswerBench+6.3points ahead
Every test, side by side
All 68 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4.8
- Terminal-Bench 2.1Claude Opus 4.8 by 20.684.664+20.6
- SWE-bench ProClaude Opus 4.8 by 13.869.255.4+13.8
- Terminal-Bench HardClaude Opus 4.8 by 12.158.346.2+12.1
- LiveBench · CodingClaude Opus 4.8 by 11.881.870+11.8
- LMArena · WebDevClaude Opus 4.8 by 93 rating points15561463+93 rating
- SWE-bench MultilingualClaude Opus 4.8 by 8.284.476.2+8.2
- SWE-bench VerifiedClaude Opus 4.8 by 888.680.6+8
- LiveBench · Agentic CodingClaude Opus 4.8 by 7.950.542.6+7.9
- SciCodeClaude Opus 4.8 by 3.654.450.8+3.6
AgentsClaude Opus 4.8
- AA ApexAgentsClaude Opus 4.8 by 15.139.424.3+15.1
- GDPValClaude Opus 4.8 by 14.947.832.9+14.9
- MCP AtlasClaude Opus 4.8 by 8.682.273.6+8.6
- Terminal-Bench 4.0Claude Opus 4.8 by 7.121.714.6+7.1
- τ-Bench V3 · BankingClaude Opus 4.8 by 4.134.230.1+4.1
- BrowseComptie84.383.4tie
ReasoningClaude Opus 4.8
- SimpleBenchClaude Opus 4.8 by 13.964.850.9+13.9
- Humanity's Last ExamClaude Opus 4.8 by 11.248.737.5+11.2
- CritPtClaude Opus 4.8 by 820.912.9+8
- LiveBench · ReasoningClaude Opus 4.8 by 6.589.282.7+6.5
- GPQA DiamondClaude Opus 4.8 by 3.29288.8+3.2
FactsClaude Opus 4.8
- AA-Omniscience · Non-hallucinationClaude Opus 4.8 by 54.860.75.9+54.8
- SimpleQA VerifiedClaude Opus 4.8 by 6.85346.2+6.8
- AA-Omniscience · AccuracyClaude Opus 4.8 by 5.848.843+5.8
MathClaude Opus 4.8
- FrontierMath Tier 4Claude Opus 4.8 by 53.756.12.4+53.7
- USAMO 2026Claude Opus 4.8 by 3696.760.7+36
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.8 by 34.78045.3+34.7
- IMOAnswerBenchDeepSeek-V4-Pro by 6.383.589.8+6.3
- LiveBench · MathematicsClaude Opus 4.8 by 3.694.390.7+3.6
- AIME 2026Claude Opus 4.8 by 3.310096.7+3.3
- HMMT Feb. 2026tie95.595.2tie
Following instructionsEven
- IFBenchDeepSeek-V4-Pro by 14.362.276.5+14.3
- LiveBench · Instruction FollowingClaude Opus 4.8 by 9.67262.4+9.6
Other results35 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 4.8 by 45.25812.8+45.2
- DSBench-HardClaude Opus 4.8 by 40.671.731.1+40.6
- AA-OmniscienceClaude Opus 4.8 by 39.528.8-10.7+39.5
- GDPval-AA (Elo)Claude Opus 4.8 by 336 rating points18901554+336 rating
- NL2RepoClaude Opus 4.8 by 31.269.738.5+31.2
- DSBench-FullStackClaude Opus 4.8 by 29.871.641.8+29.8
- GDPval-AA v2Claude Opus 4.8 by 287 rating points15931306+287 rating
- Program BenchClaude Opus 4.8 by 24.171.947.8+24.1
- Toolathlon VerifiedClaude Opus 4.8 by 2479.955.9+24
- AA Agentic IndexClaude Opus 4.8 by 14.942.627.7+14.9
- Artificial Analysis Coding IndexClaude Opus 4.8 by 14.974.359.4+14.9
- AutomationBench PublicClaude Opus 4.8 by 14.427.212.8+14.4
- AA IntelligenceClaude Opus 4.8 by 11.441.830.4+11.4
- Terminal-Bench 4.0Claude Opus 4.8 by 11.223.612.4+11.2
- HLE (wo / w tools)Claude Opus 4.8 by 9.757.948.2+9.7
- Terminal-bench 3.0Claude Opus 4.8 by 9.321.111.8+9.3
- livebench_data_analysisDeepSeek-V4-Pro by 8.56674.5+8.5
- AgentWorldBench - MCPDeepSeek-V4-Pro by 8.454.963.3+8.4
- ToolathlonClaude Opus 4.8 by 8.159.951.8+8.1
- AgentWorldBench - TerminalClaude Opus 4.8 by 7.959.251.3+7.9
- AgentWorldBench - SearchClaude Opus 4.8 by 7.535.127.6+7.5
- Tool-DecathlonClaude Opus 4.8 by 7.159.952.8+7.1
- Terminal-Bench 2.0Claude Opus 4.8 by 6.774.667.9+6.7
- AgentWorldBench - AndroidClaude Opus 4.8 by 6.361.555.2+6.3
- AgentWorldBench - SWEClaude Opus 4.8 by 4.764.159.4+4.7
- CyberGymDeepSeek-V4-Pro by 4.578.883.3+4.5
- AgentWorldBench - WebClaude Opus 4.8 by 4.454.750.3+4.4
- AgentWorldBench - OverallClaude Opus 4.8 by 3.656.653+3.6
- LiveBenchClaude Opus 4.8 by 3.677.273.6+3.6
- AgentWorldBench - OSClaude Opus 4.8 by 2.966.663.7+2.9
- HLE (with tools)DeepSeek-V4-Pro by 2.157.960+2.1
- HMMT Nov. 2025Claude Opus 4.8 by 2.196.594.4+2.1
- τ²-Bench Telecom (AA run)DeepSeek-V4-Pro by 1.894.496.2+1.8
- livebench_languageClaude Opus 4.8 by 1.679.778.1+1.6
- Agents' Last Examtie25.725.7tie
Questions people ask
Which is better, Claude Opus 4.8 or DeepSeek-V4-Pro?
Claude Opus 4.8 wins six of the seven areas where both have results: coding, agents, reasoning, facts, math and long documents. DeepSeek-V4-Pro wins none, but costs 96% less. They are level on following instructions.
Which is better for coding?
Claude Opus 4.8. It wins 9 of the 9 coding tests both models report; DeepSeek-V4-Pro wins none.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; DeepSeek-V4-Pro costs $0.43 and $0.87. That makes DeepSeek-V4-Pro about 96% cheaper for the same work.
How do you compare the two?
We use the 68 benchmark tests both models have published scores on. The verdict counts the 33 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 35 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.