Kimi K2 Instruct vs o3
Wins 0 of 6 areas
—
Wins 4 of 6 areas
Coding · Reasoning · Long documents · Following instructions
o3 is the stronger all-rounder.Kimi K2 Instruct is cheaper.
Scores updated · 33 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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 software03o33 of 3 tests
- ReasoningHard problems that need careful thinking03o33 of 3 tests
- Following instructionsDoing exactly what it is asked02o32 of 2 tests
- Long documentsFinding answers in very long texts01o31 of 1 test
- AgentsCarrying out multi-step tasks on its own11Even1 each
- FactsGetting facts right instead of making them up11Even1 each
Long documents rests on a single test.
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.
Kimi K2 Instruct costs 69% 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 o3 pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+35.6points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+29.7points ahead
- Reasons across sets of long documentsAA-LCR+21.7points ahead
Where Kimi K2 Instruct pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+18.4points ahead
- Real work tasks from 44 professionsGDPVal+5.4points ahead
Every test, side by side
All 33 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Codingo3
- Terminal-Bench Hardo3 by 13.623.537.1+13.6
- SciCodeo3 by 10.330.741+10.3
- SWE-bench Verifiedo3 by 3.365.869.1+3.3
AgentsEven
- BrowseCompo3 by 35.614.149.7+35.6
- GDPValKimi K2 Instruct by 5.418.212.8+5.4
Reasoningo3
- Humanity's Last Examo3 by 13.76.420.1+13.7
- GPQA Diamondo3 by 676.782.7+6
- CritPto3 by 1.101.1+1.1
FactsEven
- AA-Omniscience · Non-hallucinationKimi K2 Instruct by 18.430.311.9+18.4
- AA-Omniscience · Accuracyo3 by 13.225.438.6+13.2
Following instructionso3
- IFBencho3 by 29.741.771.4+29.7
- Multi-Challengeo3 by 6.354.160.4+6.3
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.
- AIME 2025o3 by 36.949.586.4+36.9
- LiveCodeBencho3 by 3153.784.7+31
- AIME 2024o3 by 2269.691.6+22
- Aider-Polygloto3 by 21.36081.3+21.3
- SimpleQAo3 by 18.43149.4+18.4
- Artificial Analysis Coding Indexo3 by 12.525.938.4+12.5
- AA-Omniscienceo3 by 11.1-26.6-15.6+11.1
- AIR-Bench 2024o3 by 10.474.184.5+10.4
- τ²-Bench (Retail)o3 by 9.670.680.2+9.6
- Tau2 airlineo3 by 8.356.564.8+8.3
- τ²-Bench Telecom (AA run)o3 by 7.373.480.7+7.3
- ZebraLogico3 by 6.88995.8+6.8
- AA Intelligenceo3 by 4.915.320.2+4.9
- MMLU-Proo3 by 3.981.185+3.9
- AA Agentic IndexKimi K2 Instruct by 1.637.736.1+1.6
- anthropic_red_teamKimi K2 Instruct by 199.398.3+1
- HarmBencho3 by 197.498.4+1
- simple_safety_testsKimi K2 Instruct by 110099+1
- XSTesttie98.297.3tie
- MATH-500 (EM)tie97.498.1tie
Questions people ask
Which is better, Kimi K2 Instruct or o3?
o3 wins four of the six areas where both have results: coding, reasoning, long documents and following instructions. Kimi K2 Instruct wins none, but costs 69% less. They are level on agents and facts.
Which is better for coding?
o3. It wins 3 of the 3 coding tests both models report; Kimi K2 Instruct wins none.
Which is cheaper?
Kimi K2 Instruct costs $0.60 per million input tokens and $2.50 per million output tokens; o3 costs $2.00 and $8.00. That makes Kimi K2 Instruct about 69% cheaper for the same work.
How do you compare the two?
We use the 33 benchmark tests both models have published scores on. The verdict counts the 13 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.