GPT-5.4 vs Kimi K3
Wins 1 of 8 areas
Math
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents · Following instructions
Kimi K3 is the stronger all-rounder.GPT-5.4 is cheaper and better at math.
Scores updated · 46 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- AgentsCarrying out multi-step tasks on its own07Kimi K37 of 7 tests
- CodingWriting and fixing software05Kimi K35 of 5 tests
- ReasoningHard problems that need careful thinking13Kimi K33 of 5 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video02Kimi K32 of 2 tests
- FactsGetting facts right instead of making them up12Kimi 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
- MathCompetition and research-level math40GPT-5.44 of 5 tests · 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.
GPT-5.4 costs 3% 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+38.5points ahead
- Multi-step tasks using many real software toolsMCP Atlas+17points ahead
- Real work tasks from 44 professionsGDPVal+14.4points ahead
Where GPT-5.4 pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+12.9points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+10points ahead
- Fresh competition math problemsLiveBench · Mathematics+9.8points ahead
Every test, side by side
All 46 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingKimi K3
- LMArena · WebDevKimi K3 by 193 rating points14651658+193 rating
- LiveBench · Agentic CodingKimi K3 by 8.453.862.2+8.4
- Terminal-Bench 2.1Kimi K3 by 6.778.385+6.7
- LiveBench · CodingKimi K3 by 477.581.5+4
- SciCodeKimi K3 by 2.956.659.5+2.9
AgentsKimi K3
- MCP AtlasKimi K3 by 1767.284.2+17
- GDPValKimi K3 by 14.437.451.8+14.4
- AA IT-Bench SREKimi K3 by 13.234.547.7+13.2
- OSWorld-VerifiedKimi K3 by 9.87584.8+9.8
- BrowseCompKimi K3 by 8.582.791.2+8.5
- AA ApexAgentsKimi K3 by 833.341.3+8
- τ-Bench V3 · BankingKimi K3 by 6.439.646+6.4
ReasoningKimi K3
- ARC-AGI-2GPT-5.4 by 12.973.360.4+12.9
- Humanity's Last ExamKimi K3 by 3.243.746.9+3.2
- LiveBench · ReasoningKimi K3 by 2.688.190.7+2.6
- GPQA DiamondKimi K3 by 1.59293.5+1.5
- CritPttie23.423.4tie
FactsKimi K3
- AA-Omniscience · Non-hallucinationKimi K3 by 38.58.346.8+38.5
- SimpleQA VerifiedKimi K3 by 5.545.150.6+5.5
- AA-Omniscience · AccuracyGPT-5.4 by 3.350.947.6+3.3
Images and chartsKimi K3
- CharXiv (RQ)Kimi K3 by 8.582.891.3+8.5
- MMMU-ProKimi K3 by 2.178.480.5+2.1
MathGPT-5.4
- FrontierMath Tier 4GPT-5.4 by 104939+10
- LiveBench · MathematicsGPT-5.4 by 9.894.284.4+9.8
- FrontierMath Tiers 1-3 (v2)GPT-5.4 by 6.478.672.2+6.4
- AIME 2026GPT-5.4 by 2.599.296.7+2.5
- HMMT Feb. 2026tie97.797tie
Following instructionsKimi K3
- LiveBench · Instruction FollowingKimi K3 by 1.270.271.4+1.2
Other results17 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.
- BabyVisionKimi K3 by 3649.785.7+36
- ToolathlonKimi K3 by 18.654.673.2+18.6
- DeepSWE 1.1Kimi K3 by 175269+17
- NL2RepoKimi K3 by 16.741.358+16.7
- DeepSearchQA (F1)Kimi K3 by 16.478.695+16.4
- AA-OmniscienceKimi K3 by 13.95.819.7+13.9
- CyberGymKimi K3 by 13.766.380+13.7
- OfficeQA ProKimi K3 by 12.251.163.3+12.2
- HLE (with tools)Kimi K3 by 7.85259.8+7.8
- AA Agentic IndexKimi K3 by 6.444.250.6+6.4
- MATH-VisionKimi K3 by 5.89297.8+5.8
- Artificial Analysis Coding IndexKimi K3 by 5.171.176.2+5.1
- AA IntelligenceKimi K3 by 4.63943.6+4.6
- livebench_languageKimi K3 by 2.982.685.5+2.9
- matharena_visual_math_overallGPT-5.4 by 1.292.591.3+1.2
- ARC-AGI-1tie93.794.5tie
- livebench_data_analysistie79.378.7tie
Questions people ask
Which is better, GPT-5.4 or Kimi K3?
Kimi K3 wins seven of the eight areas we test: coding, agents, reasoning, facts, images and charts, long documents and following instructions. GPT-5.4 wins math, and costs 3% less.
Which is better for coding?
Kimi K3. It wins 5 of the 5 coding tests both models report; GPT-5.4 wins none.
Which is cheaper?
GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens; Kimi K3 costs $3.00 and $15.00. That makes GPT-5.4 about 3% cheaper for the same work.
How do you compare the two?
We use the 46 benchmark tests both models have published scores on. The verdict counts the 29 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 17 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.