Kimi K2.6 vs Qwen3.7 Max
Wins 1 of 7 areas
Long documents
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Following instructions
Qwen3.7 Max is the stronger all-rounder.Kimi K2.6 is cheaper and better at long documents.
Scores updated · 46 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.
- MathCompetition and research-level math05Qwen3.7 Max5 of 6 tests · 1 tie
- ReasoningHard problems that need careful thinking04Qwen3.7 Max4 of 4 tests
- AgentsCarrying out multi-step tasks on its own14Qwen3.7 Max4 of 5 tests
- CodingWriting and fixing software35Qwen3.7 Max5 of 10 tests · 2 ties
- Following instructionsDoing exactly what it is asked02Qwen3.7 Max2 of 2 tests
- FactsGetting facts right instead of making them up12Qwen3.7 Max2 of 3 tests
- Long documentsFinding answers in very long texts10Kimi K2.61 of 1 test
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 K2.6 costs 51% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.7 Max pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+20.9points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+14.9points ahead
- Finds the root cause of IT system incidentsAA IT-Bench SRE+11.3points ahead
Where Kimi K2.6 pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+11.5points ahead
- Programming problems, refreshed regularlyLiveBench · Coding+4.4points ahead
- Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+3.3points 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.
CodingQwen3.7 Max
- Terminal-Bench 2.1Qwen3.7 Max by 8.665.974.5+8.6
- Terminal-Bench HardQwen3.7 Max by 6.943.950.8+6.9
- LiveBench · CodingKimi K2.6 by 4.478.674.2+4.4
- LiveBench · Agentic CodingKimi K2.6 by 3.346.943.6+3.3
- LiveCodeBench v6Qwen3.7 Max by 289.691.6+2
- SciCodeKimi K2.6 by 251.549.5+2
- SWE-bench ProQwen3.7 Max by 258.660.6+2
- SWE-bench MultilingualQwen3.7 Max by 1.676.778.3+1.6
- LMArena · WebDevtie15091515tie
- SWE-bench Verifiedtie80.280.4tie
AgentsQwen3.7 Max
- τ-Bench V3 · BankingKimi K2.6 by 11.523.311.8+11.5
- AA IT-Bench SREQwen3.7 Max by 11.331.242.5+11.3
- MCP AtlasQwen3.7 Max by 8.368.176.4+8.3
- GDPValQwen3.7 Max by 4.62731.6+4.6
- Terminal-Bench 4.0Qwen3.7 Max by 10.51.5+1
ReasoningQwen3.7 Max
- CritPtQwen3.7 Max by 5.4813.4+5.4
- LiveBench · ReasoningQwen3.7 Max by 3.979.483.3+3.9
- Humanity's Last ExamQwen3.7 Max by 337.540.5+3
- GPQA DiamondQwen3.7 Max by 1.291.192.3+1.2
FactsQwen3.7 Max
- SimpleQA VerifiedQwen3.7 Max by 20.934.955.8+20.9
- AA-Omniscience · Non-hallucinationQwen3.7 Max by 14.959.574.4+14.9
- AA-Omniscience · AccuracyKimi K2.6 by 1.532.631.1+1.5
MathQwen3.7 Max
- FrontierMath Tier 4Qwen3.7 Max by 8.625.634.1+8.6
- FrontierMath Tiers 1-3 (v2)Qwen3.7 Max by 7.457.264.6+7.4
- HMMT Feb. 2026Qwen3.7 Max by 4.492.797.1+4.4
- IMOAnswerBenchQwen3.7 Max by 48690+4
- LiveBench · MathematicsQwen3.7 Max by 184.385.3+1
- AIME 2026tie96.497tie
Following instructionsQwen3.7 Max
- LiveBench · Instruction FollowingQwen3.7 Max by 9.664.474+9.6
- IFBenchQwen3.7 Max by 4.57680.5+4.5
Other results15 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.
- Claw Eval (pass@3)Kimi K2.6 by 15.780.965.2+15.7
- AA-OmniscienceQwen3.7 Max by 8.25.313.5+8.2
- livebench_data_analysisQwen3.7 Max by 6.765.171.8+6.7
- WideSearchKimi K2.6 by 5.680.875.2+5.6
- MCPMarkQwen3.7 Max by 4.955.960.8+4.9
- livebench_languageQwen3.7 Max by 4.675.179.7+4.6
- Artificial Analysis Coding IndexQwen3.7 Max by 4.261.866+4.2
- Finance Agent v2Qwen3.7 Max by 3.544.948.4+3.5
- Terminal-Bench 2.0Qwen3.7 Max by 366.769.7+3
- AA IntelligenceQwen3.7 Max by 2.52729.5+2.5
- MMLU-ProQwen3.7 Max by 2.587.189.6+2.5
- LiveBenchQwen3.7 Max by 2.172.274.3+2.1
- AA Agentic IndexQwen3.7 Max by 1.822.123.9+1.8
- τ²-Bench Telecom (AA run)Kimi K2.6 by 1.295.994.7+1.2
- HLE (with tools)tie5453.5tie
Questions people ask
Which is better, Kimi K2.6 or Qwen3.7 Max?
Qwen3.7 Max wins six of the seven areas where both have results: coding, agents, reasoning, facts, math and following instructions. Kimi K2.6 wins long documents, and costs 51% less.
Which is better for coding?
Qwen3.7 Max. It wins 5 of the 10 coding tests both models report; Kimi K2.6 wins 3, and 2 are ties.
Which is cheaper?
Kimi K2.6 costs $0.95 per million input tokens and $4.00 per million output tokens; Qwen3.7 Max costs $2.50 and $7.50. That makes Kimi K2.6 about 51% 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 31 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 15 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.