GPT-5.6 Luna vs Kimi K2.6
Wins 6 of 8 areas
Coding · Agents · Reasoning · Facts · Math · Long documents
Wins 2 of 8 areas
Images and charts · Following instructions
GPT-5.6 Luna is the stronger all-rounder.Kimi K2.6 is better at images and charts.
Scores updated · 42 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.
- CodingWriting and fixing software70GPT-5.6 Luna7 of 7 tests
- MathCompetition and research-level math50GPT-5.6 Luna5 of 5 tests
- AgentsCarrying out multi-step tasks on its own51GPT-5.6 Luna5 of 7 tests · 1 tie
- ReasoningHard problems that need careful thinking30GPT-5.6 Luna3 of 4 tests · 1 tie
- FactsGetting facts right instead of making them up21GPT-5.6 Luna2 of 3 tests
- Long documentsFinding answers in very long texts10GPT-5.6 Luna1 of 1 test
- Images and chartsUnderstanding pictures, charts and video02Kimi K2.62 of 3 tests · 1 tie
- Following instructionsDoing exactly what it is asked01Kimi K2.61 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.
GPT-5.6 Luna costs 72% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GPT-5.6 Luna pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+32.9points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+24.9points ahead
- Real work tasks from 44 professionsGDPVal+21.1points ahead
Where Kimi K2.6 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+52.1points ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+4.3points ahead
- Reasoning about charts from research papersCharXiv (RQ)+4points ahead
Every test, side by side
All 42 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.6 Luna
- Terminal-Bench 2.1GPT-5.6 Luna by 1580.965.9+15
- SWE-bench VerifiedGPT-5.6 Luna by 12.89380.2+12.8
- LiveBench · CodingGPT-5.6 Luna by 4.382.978.6+4.3
- SWE-bench ProGPT-5.6 Luna by 4.162.758.6+4.1
- SciCodeGPT-5.6 Luna by 2.153.651.5+2.1
- LiveBench · Agentic CodingGPT-5.6 Luna by 1.548.446.9+1.5
- LMArena · WebDevGPT-5.6 Luna by 10 rating points15191509+10 rating
AgentsGPT-5.6 Luna
- GDPValGPT-5.6 Luna by 21.14826.9+21.1
- Terminal-Bench 4.0GPT-5.6 Luna by 11.111.60.5+11.1
- AA IT-Bench SREGPT-5.6 Luna by 9.140.331.2+9.1
- τ-Bench V3 · BankingGPT-5.6 Luna by 7.831.123.3+7.8
- AA ApexAgentsGPT-5.6 Luna by 7.335.828.5+7.3
- BrowseCompKimi K2.6 by 383.386.3+3
- OSWorld-Verifiedtie72.673.1tie
ReasoningGPT-5.6 Luna
- CritPtGPT-5.6 Luna by 12.620.68+12.6
- LiveBench · ReasoningGPT-5.6 Luna by 6.285.679.4+6.2
- Humanity's Last ExamGPT-5.6 Luna by 239.537.5+2
- GPQA Diamondtie91.191.1tie
FactsGPT-5.6 Luna
- AA-Omniscience · Non-hallucinationKimi K2.6 by 52.17.459.5+52.1
- AA-Omniscience · AccuracyGPT-5.6 Luna by 10.142.732.6+10.1
- SimpleQA VerifiedGPT-5.6 Luna by 6.14134.9+6.1
Images and chartsKimi K2.6
- CharXiv (RQ)Kimi K2.6 by 482.786.7+4
- LMArena · VisionKimi K2.6 by 26 rating points12561282+26 rating
- MMMU-Protie78.679.4tie
MathGPT-5.6 Luna
- FrontierMath Tier 4GPT-5.6 Luna by 32.958.525.6+32.9
- FrontierMath Tiers 1-3 (v2)GPT-5.6 Luna by 24.982.157.2+24.9
- HMMT Feb. 2026GPT-5.6 Luna by 5.898.592.7+5.8
- LiveBench · MathematicsGPT-5.6 Luna by 2.987.284.3+2.9
- AIME 2026GPT-5.6 Luna by 1.297.696.4+1.2
Following instructionsKimi K2.6
- LiveBench · Instruction FollowingKimi K2.6 by 4.360.164.4+4.3
Other results11 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.
- GDPval-AA v2GPT-5.6 Luna by 340 rating points15301190+340 rating
- AA Agentic IndexGPT-5.6 Luna by 20.642.722.1+20.6
- AA-OmniscienceKimi K2.6 by 15.6-10.35.3+15.6
- livebench_data_analysisGPT-5.6 Luna by 12.97865.1+12.9
- AA IntelligenceGPT-5.6 Luna by 10.337.327+10.3
- Artificial Analysis Coding IndexGPT-5.6 Luna by 9.671.461.8+9.6
- HLE (with tools)Kimi K2.6 by 5.148.954+5.1
- SWEBench Pro PublicGPT-5.6 Luna by 4.162.758.6+4.1
- τ³-Bench BankingGPT-5.6 Luna by 3.724.320.6+3.7
- ToolathlonGPT-5.6 Luna by 3.453.450+3.4
- livebench_languageKimi K2.6 by 2.572.675.1+2.5
Questions people ask
Which is better, GPT-5.6 Luna or Kimi K2.6?
GPT-5.6 Luna wins six of the eight areas we test: coding, agents, reasoning, facts, math and long documents. Kimi K2.6 wins images and charts and following instructions.
Which is better for coding?
GPT-5.6 Luna. It wins 7 of the 7 coding tests both models report; Kimi K2.6 wins none.
Which is cheaper?
GPT-5.6 Luna costs $0.20 per million input tokens and $1.20 per million output tokens; Kimi K2.6 costs $0.95 and $4.00. That makes GPT-5.6 Luna about 72% cheaper for the same work.
How do you compare the two?
We use the 42 benchmark tests both models have published scores on. The verdict counts the 31 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 11 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.