GPT-5.2 vs Kimi K3
Wins 0 of 8 areas
—
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents · Following instructions
Kimi K3 is the stronger all-rounder.GPT-5.2 is cheaper.
Scores updated · 41 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.
- ReasoningHard problems that need careful thinking06Kimi K36 of 6 tests
- AgentsCarrying out multi-step tasks on its own05Kimi K35 of 5 tests
- CodingWriting and fixing software04Kimi K34 of 4 tests
- FactsGetting facts right instead of making them up03Kimi K33 of 3 tests
- Images and chartsUnderstanding pictures, charts and video02Kimi K32 of 2 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 math22Even2 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.
GPT-5.2 costs 13% 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 Kimi K3 pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+34.9points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+28points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+25.4points ahead
Where GPT-5.2 pulls ahead
- Fresh competition math problemsLiveBench · Mathematics+8.8points ahead
- US invitational high-school math exam problemsAIME 2026+1.6points 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.
CodingKimi K3
- LMArena · WebDevKimi K3 by 239 rating points14161655+239 rating
- LiveBench · Agentic CodingKimi K3 by 11.950.362.2+11.9
- SciCodeKimi K3 by 7.452.159.5+7.4
- LiveBench · CodingKimi K3 by 5.476.181.5+5.4
AgentsKimi K3
- τ-Bench V3 · BankingKimi K3 by 34.911.146+34.9
- BrowseCompKimi K3 by 25.465.891.2+25.4
- MCP AtlasKimi K3 by 23.660.684.2+23.6
- AA ApexAgentsKimi K3 by 18.32341.3+18.3
- GDPValKimi K3 by 3.448.351.7+3.4
ReasoningKimi K3
- SimpleBenchKimi K3 by 14.945.860.7+14.9
- CritPtKimi K3 by 11.811.623.4+11.8
- Humanity's Last ExamKimi K3 by 9.237.746.9+9.2
- ARC-AGI-2Kimi K3 by 7.552.960.4+7.5
- LiveBench · ReasoningKimi K3 by 7.583.290.7+7.5
- GPQA DiamondKimi K3 by 3.290.393.5+3.2
FactsKimi K3
- AA-Omniscience · Non-hallucinationKimi K3 by 2818.846.8+28
- SimpleQA VerifiedKimi K3 by 13.537.150.6+13.5
- AA-Omniscience · AccuracyKimi K3 by 3.344.347.6+3.3
Images and chartsKimi K3
- CharXiv (RQ)Kimi K3 by 9.282.191.3+9.2
- MMMU-ProKimi K3 by 5.974.680.5+5.9
MathEven
- LiveBench · MathematicsGPT-5.2 by 8.893.284.4+8.8
- FrontierMath Tier 4Kimi K3 by 7.331.739+7.3
- FrontierMath Tiers 1-3 (v2)Kimi K3 by 4.867.472.2+4.8
- AIME 2026GPT-5.2 by 1.698.396.7+1.6
- HMMT Feb. 2026tie9797tie
Following instructionsKimi K3
- LiveBench · Instruction FollowingKimi K3 by 9.661.871.4+9.6
Other results14 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.
- ZeroBenchKimi K3 by 32941+32
- Artificial Analysis Coding IndexKimi K3 by 27.548.776.2+27.5
- ToolathlonKimi K3 by 26.946.373.2+26.9
- DeepSearchQA (F1)Kimi K3 by 23.771.395+23.7
- WorldVQAKimi K3 by 232851+23
- AA-OmniscienceKimi K3 by 20.6-0.919.7+20.6
- MATH-VisionKimi K3 by 14.88397.8+14.8
- HLE (with tools)Kimi K3 by 14.345.559.8+14.3
- AA IntelligenceKimi K3 by 13.230.443.6+13.2
- AA Agentic IndexGPT-5.2 by 9.660.250.6+9.6
- ARC-AGI-1Kimi K3 by 8.386.294.5+8.3
- livebench_languageKimi K3 by 5.779.885.5+5.7
- matharena_visual_math_overallKimi K3 by 4.886.591.3+4.8
- livebench_data_analysistie78.278.7tie
Questions people ask
Which is better, GPT-5.2 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-5.2 wins none, but costs 13% less. They are level on math.
Which is better for coding?
Kimi K3. It wins 4 of the 4 coding tests both models report; GPT-5.2 wins none.
Which is cheaper?
GPT-5.2 costs $1.75 per million input tokens and $14.00 per million output tokens; Kimi K3 costs $3.00 and $15.00. That makes GPT-5.2 about 13% 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 27 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 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.