DeepSeek-V4-Pro 0813 vs Kimi K3
Wins 1 of 7 areas
Facts
Wins 6 of 7 areas
Coding · Agents · Reasoning · Math · Long documents · Following instructions
Kimi K3 is the stronger all-rounder.DeepSeek-V4-Pro 0813 is cheaper and better at facts.
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.
- CodingWriting and fixing software04Kimi K34 of 4 tests
- ReasoningHard problems that need careful thinking03Kimi K33 of 5 tests · 2 ties
- AgentsCarrying out multi-step tasks on its own12Kimi K32 of 3 tests
- MathCompetition and research-level math12Kimi K32 of 3 tests
- Long documentsFinding answers in very long texts01Kimi K31 of 1 test
- Following instructionsDoing exactly what it is asked01Kimi K31 of 1 test
- FactsGetting facts right instead of making them up21DeepSeek-V4-Pro 08132 of 3 tests
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 71% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Kimi K3 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+41.6points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+12.2points ahead
- Code for real scientific research problemsSciCode+8.5points ahead
Where DeepSeek-V4-Pro 0813 pulls ahead
- Fresh competition math problemsLiveBench · Mathematics+10.7points ahead
- Short factual questions, answered correctlySimpleQA Verified+2.3points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+1.5points 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.
CodingKimi K3
- SciCodeKimi K3 by 8.55159.5+8.5
- LiveBench · Agentic CodingKimi K3 by 7.25562.2+7.2
- Terminal-Bench 2.1Kimi K3 by 6.378.785+6.3
- LiveBench · CodingKimi K3 by 4.377.281.5+4.3
AgentsKimi K3
- τ-Bench V3 · BankingKimi K3 by 6.439.646+6.4
- GDPValKimi K3 by 4.247.551.7+4.2
- Terminal-Bench 4.0DeepSeek-V4-Pro 0813 by 1.514.112.6+1.5
ReasoningKimi K3
- Humanity's Last ExamKimi K3 by 5.94146.9+5.9
- CritPtKimi K3 by 5.41823.4+5.4
- LiveBench · ReasoningKimi K3 by 4.985.890.7+4.9
- ARC-AGI-2tie61.360.4tie
- GPQA Diamondtie92.893.5tie
FactsDeepSeek-V4-Pro 0813
- AA-Omniscience · Non-hallucinationKimi K3 by 41.65.246.8+41.6
- SimpleQA VerifiedDeepSeek-V4-Pro 0813 by 2.352.950.6+2.3
- AA-Omniscience · AccuracyDeepSeek-V4-Pro 0813 by 1.549.147.6+1.5
MathKimi K3
- FrontierMath Tier 4Kimi K3 by 12.226.839+12.2
- LiveBench · MathematicsDeepSeek-V4-Pro 0813 by 10.795.184.4+10.7
- FrontierMath Tiers 1-3 (v2)Kimi K3 by 7.664.672.2+7.6
Following instructionsKimi K3
- LiveBench · Instruction FollowingKimi K3 by 3.767.771.4+3.7
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-OmniscienceKimi K3 by 18.90.819.7+18.9
- GDPval-AA v2Kimi K3 by 92 rating points15901682+92 rating
- AA Agentic IndexKimi K3 by 8.342.350.6+8.3
- AA IntelligenceKimi K3 by 7.63643.6+7.6
- Artificial Analysis Coding IndexKimi K3 by 7.468.876.2+7.4
- DeepSWEKimi K3 by 4.862.767.5+4.8
- ARC-AGI-1Kimi K3 by 4.59094.5+4.5
- DSBench-HardDeepSeek-V4-Pro 0813 by 4.267.263+4.2
- HLE (wo / w tools)DeepSeek-V4-Pro 0813 by 46056+4
- NL2RepoDeepSeek-V4-Pro 0813 by 3.561.558+3.5
- livebench_languageKimi K3 by 3.482.185.5+3.4
- CyberGymDeepSeek-V4-Pro 0813 by 3.383.380+3.3
- Agents' Last ExamKimi K3 by 2.625.728.3+2.6
- DSBench-FullStackKimi K3 by 2.671.173.7+2.6
- Toolathlon VerifiedKimi K3 by 2.474.176.5+2.4
- Agents' Last Exam (ALE-CLI)Kimi K3 by 1.925.727.6+1.9
- AutomationBench PublicDeepSeek-V4-Pro 0813 by 131.830.8+1
- Toolathlontie74.173.2tie
- livebench_data_analysistie79.278.7tie
- HLE (with tools)tie6059.8tie
Questions people ask
Which is better, DeepSeek-V4-Pro 0813 or Kimi K3?
Kimi K3 wins six of the seven areas where both have results: coding, agents, reasoning, math, long documents and following instructions. DeepSeek-V4-Pro 0813 wins facts, and costs 71% less.
Which is better for coding?
Kimi K3. It wins 4 of the 4 coding tests both models report; DeepSeek-V4-Pro 0813 wins none.
Which is cheaper?
DeepSeek-V4-Pro 0813 costs $1.32 per million input tokens and $3.96 per million output tokens; Kimi K3 costs $3.00 and $15.00. That makes DeepSeek-V4-Pro 0813 about 71% 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.