Claude Sonnet 4.5 vs DeepSeek-V4-Pro
Wins 0 of 7 areas
—
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Long documents · Following instructions
DeepSeek-V4-Pro is the stronger all-rounder.
Scores updated · 41 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 software08DeepSeek-V4-Pro8 of 8 tests
- AgentsCarrying out multi-step tasks on its own05DeepSeek-V4-Pro5 of 5 tests
- ReasoningHard problems that need careful thinking13DeepSeek-V4-Pro3 of 4 tests
- FactsGetting facts right instead of making them up13DeepSeek-V4-Pro3 of 4 tests
- MathCompetition and research-level math02DeepSeek-V4-Pro2 of 3 tests · 1 tie
- Long documentsFinding answers in very long texts01DeepSeek-V4-Pro1 of 1 test
- Following instructionsDoing exactly what it is asked01DeepSeek-V4-Pro1 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 93% 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 DeepSeek-V4-Pro pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+59.3points ahead
- Multi-step tasks using many real software toolsMCP Atlas+29.8points ahead
- Recent programming contest problemsLiveCodeBench v6+28.5points ahead
Where Claude Sonnet 4.5 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+45points ahead
- Common-sense trick questionsSimpleBench+3.4points ahead
Every test, side by side
All 41 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-V4-Pro
- LiveCodeBench v6DeepSeek-V4-Pro by 28.56492.5+28.5
- SWE-bench ProDeepSeek-V4-Pro by 11.843.655.4+11.8
- Terminal-Bench HardDeepSeek-V4-Pro by 10.635.646.2+10.6
- SWE-bench MultilingualDeepSeek-V4-Pro by 9.26776.2+9.2
- Terminal-Bench 2.1DeepSeek-V4-Pro by 8.255.864+8.2
- LMArena · WebDevDeepSeek-V4-Pro by 70 rating points13941464+70 rating
- SciCodeDeepSeek-V4-Pro by 5.145.750.8+5.1
- SWE-bench VerifiedDeepSeek-V4-Pro by 3.477.280.6+3.4
AgentsDeepSeek-V4-Pro
- BrowseCompDeepSeek-V4-Pro by 59.324.183.4+59.3
- MCP AtlasDeepSeek-V4-Pro by 29.843.873.6+29.8
- Terminal-Bench 4.0DeepSeek-V4-Pro by 14.6014.6+14.6
- GDPValDeepSeek-V4-Pro by 12.420.532.9+12.4
- τ-Bench V3 · BankingDeepSeek-V4-Pro by 5.624.530.1+5.6
ReasoningDeepSeek-V4-Pro
- Humanity's Last ExamDeepSeek-V4-Pro by 19.717.837.5+19.7
- CritPtDeepSeek-V4-Pro by 11.81.112.9+11.8
- GPQA DiamondDeepSeek-V4-Pro by 5.483.488.8+5.4
- SimpleBenchClaude Sonnet 4.5 by 3.454.350.9+3.4
FactsDeepSeek-V4-Pro
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.5 by 4550.95.9+45
- SimpleQA VerifiedDeepSeek-V4-Pro by 15.530.746.2+15.5
- AA-Omniscience · AccuracyDeepSeek-V4-Pro by 10.132.943+10.1
- Vectara HHEM hallucination ratelower is betterDeepSeek-V4-Pro by 3.4128.6+3.4
MathDeepSeek-V4-Pro
- IMOAnswerBenchDeepSeek-V4-Pro by 23.965.989.8+23.9
- FrontierMath Tiers 1-3 (v2)DeepSeek-V4-Pro by 21.423.945.3+21.4
- FrontierMath Tier 4tie2.42.4tie
Following instructionsDeepSeek-V4-Pro
- IFBenchDeepSeek-V4-Pro by 19.257.376.5+19.2
Other results15 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.
- HLE (with tools)DeepSeek-V4-Pro by 283260+28
- LiveCodeBenchDeepSeek-V4-Pro by 22.57193.5+22.5
- τ²-Bench Telecom (AA run)DeepSeek-V4-Pro by 18.178.196.2+18.1
- Terminal-Bench 2.0DeepSeek-V4-Pro by 17.95067.9+17.9
- HMMT Nov. 2025DeepSeek-V4-Pro by 12.781.794.4+12.7
- ToolathlonDeepSeek-V4-Pro by 10.84151.8+10.8
- AA-OmniscienceClaude Sonnet 4.5 by 10.6-0.1-10.7+10.6
- AA Agentic IndexDeepSeek-V4-Pro by 10.217.527.7+10.2
- AA IntelligenceDeepSeek-V4-Pro by 9.720.730.4+9.7
- Artificial Analysis Coding IndexDeepSeek-V4-Pro by 7.352.159.4+7.3
- GPQA (unspecified)DeepSeek-V4-Pro by 4.483.487.8+4.4
- vectara_factual_consistencyDeepSeek-V4-Pro by 3.48891.4+3.4
- vectara_avg_summary_lengthDeepSeek-V4-Pro by 26 rating points127.8153.8+26 rating
- vectara_answer_rateDeepSeek-V4-Pro by 1.695.697.2+1.6
- MMLU-Protie87.587.5tie
Questions people ask
Which is better, Claude Sonnet 4.5 or DeepSeek-V4-Pro?
DeepSeek-V4-Pro wins all seven areas where both have results: coding, agents, reasoning, facts, math, long documents and following instructions. Claude Sonnet 4.5 wins none.
Which is better for coding?
DeepSeek-V4-Pro. It wins 8 of the 8 coding tests both models report; Claude Sonnet 4.5 wins none.
Which is cheaper?
Claude Sonnet 4.5 costs $3.00 per million input tokens and $15.00 per million output tokens; DeepSeek-V4-Pro costs $0.43 and $0.87. That makes DeepSeek-V4-Pro about 93% cheaper for the same work.
How do you compare the two?
We use the 41 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 15 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.