Claude Opus 4.8 vs DeepSeek-V4-Pro 0813
Wins 5 of 7 areas
Coding · Reasoning · Facts · Math · Following instructions
Wins 1 of 7 areas
Long documents
Claude Opus 4.8 is the stronger all-rounder.DeepSeek-V4-Pro 0813 is cheaper and better at long documents.
Scores updated · 40 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.
- ReasoningHard problems that need careful thinking40Claude Opus 4.84 of 5 tests · 1 tie
- CodingWriting and fixing software31Claude Opus 4.83 of 4 tests
- MathCompetition and research-level math20Claude Opus 4.82 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up10Claude Opus 4.81 of 3 tests · 2 ties
- Following instructionsDoing exactly what it is asked10Claude Opus 4.81 of 1 test
- Long documentsFinding answers in very long texts01DeepSeek-V4-Pro 08131 of 1 test
- AgentsCarrying out multi-step tasks on its own11Even1 each · 1 tie
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 0813 costs 82% 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+55.5points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+29.3points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+15.4points ahead
Where DeepSeek-V4-Pro 0813 pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+5.4points ahead
- Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+4.5points ahead
- Reasons across sets of long documentsAA-LCR+2.6points ahead
Every test, side by side
All 40 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 5.984.678.7+5.9
- LiveBench · CodingClaude Opus 4.8 by 4.681.877.2+4.6
- LiveBench · Agentic CodingDeepSeek-V4-Pro 0813 by 4.550.555+4.5
- SciCodeClaude Opus 4.8 by 3.454.451+3.4
AgentsEven
- Terminal-Bench 4.0Claude Opus 4.8 by 7.621.714.1+7.6
- τ-Bench V3 · BankingDeepSeek-V4-Pro 0813 by 5.434.239.6+5.4
- GDPValtie47.747.5tie
ReasoningClaude Opus 4.8
- ARC-AGI-2Claude Opus 4.8 by 10.872.161.3+10.8
- Humanity's Last ExamClaude Opus 4.8 by 7.748.741+7.7
- LiveBench · ReasoningClaude Opus 4.8 by 3.489.285.8+3.4
- CritPtClaude Opus 4.8 by 2.920.918+2.9
- GPQA Diamondtie9292.8tie
FactsClaude Opus 4.8
- AA-Omniscience · Non-hallucinationClaude Opus 4.8 by 55.560.75.2+55.5
- AA-Omniscience · Accuracytie48.849.1tie
- SimpleQA Verifiedtie5352.9tie
MathClaude Opus 4.8
- FrontierMath Tier 4Claude Opus 4.8 by 29.356.126.8+29.3
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.8 by 15.48064.6+15.4
- LiveBench · Mathematicstie94.395.1tie
Long documentsDeepSeek-V4-Pro 0813
- AA-LCRDeepSeek-V4-Pro 0813 by 2.677.780.3+2.6
Following instructionsClaude Opus 4.8
- LiveBench · Instruction FollowingClaude Opus 4.8 by 4.37267.7+4.3
Other results20 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.
- AA-OmniscienceClaude Opus 4.8 by 2828.80.8+28
- ToolathlonDeepSeek-V4-Pro 0813 by 14.259.974.1+14.2
- livebench_data_analysisDeepSeek-V4-Pro 0813 by 13.26679.2+13.2
- NL2RepoClaude Opus 4.8 by 8.269.761.5+8.2
- AA IntelligenceClaude Opus 4.8 by 5.841.836+5.8
- Toolathlon VerifiedClaude Opus 4.8 by 5.879.974.1+5.8
- Artificial Analysis Coding IndexClaude Opus 4.8 by 5.574.368.8+5.5
- DeepSWEDeepSeek-V4-Pro 0813 by 4.75862.7+4.7
- AutomationBench PublicDeepSeek-V4-Pro 0813 by 4.627.231.8+4.6
- CyberGymDeepSeek-V4-Pro 0813 by 4.578.883.3+4.5
- DSBench-HardClaude Opus 4.8 by 4.571.767.2+4.5
- ARC-AGI-1Claude Opus 4.8 by 2.592.590+2.5
- livebench_languageDeepSeek-V4-Pro 0813 by 2.479.782.1+2.4
- HLE (with tools)DeepSeek-V4-Pro 0813 by 2.157.960+2.1
- HLE (wo / w tools)DeepSeek-V4-Pro 0813 by 2.157.960+2.1
- DSBench-FullStacktie71.671.1tie
- AA Agentic Indextie42.642.3tie
- GDPval-AA v2tie15931590tie
- Agents' Last Examtie25.725.7tie
- Agents' Last Exam (ALE-CLI)tie25.725.7tie
Questions people ask
Which is better, Claude Opus 4.8 or DeepSeek-V4-Pro 0813?
Claude Opus 4.8 wins five of the seven areas where both have results: coding, reasoning, facts, math and following instructions. DeepSeek-V4-Pro 0813 wins long documents, and costs 82% less. They are level on agents.
Which is better for coding?
Claude Opus 4.8. It wins 3 of the 4 coding tests both models report; DeepSeek-V4-Pro 0813 wins 1.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; DeepSeek-V4-Pro 0813 costs $1.32 and $3.96. That makes DeepSeek-V4-Pro 0813 about 82% cheaper for the same work.
How do you compare the two?
We use the 40 benchmark tests both models have published scores on. The verdict counts the 20 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 20 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.