Kimi K2.5 vs Qwen3.6 Plus
Wins 1 of 8 areas
Reasoning
Wins 7 of 8 areas
Coding · Agents · Facts · Images and charts · Math · Long documents · Following instructions
Qwen3.6 Plus is the stronger all-rounder.Kimi K2.5 is better at reasoning.
Scores updated · 49 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 software16Qwen3.6 Plus6 of 8 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own03Qwen3.6 Plus3 of 4 tests · 1 tie
- FactsGetting facts right instead of making them up12Qwen3.6 Plus2 of 3 tests
- Images and chartsUnderstanding pictures, charts and video12Qwen3.6 Plus2 of 3 tests
- MathCompetition and research-level math01Qwen3.6 Plus1 of 3 tests · 2 ties
- Long documentsFinding answers in very long texts01Qwen3.6 Plus1 of 2 tests · 1 tie
- Following instructionsDoing exactly what it is asked01Qwen3.6 Plus1 of 1 test
- ReasoningHard problems that need careful thinking10Kimi K2.51 of 3 tests · 2 ties
Following instructions 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.
Qwen3.6 Plus 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 Qwen3.6 Plus pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+31.1points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+15.7points ahead
- Multi-step tasks using many real software toolsMCP Atlas+10.3points ahead
Where Kimi K2.5 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+8.8points ahead
- Code for real scientific research problemsSciCode+8.3points ahead
- Questions about short and long videosVideo-MME+3.2points ahead
Every test, side by side
All 49 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.6 Plus
- Terminal-Bench 2.1Qwen3.6 Plus by 15.745.761.4+15.7
- Terminal-Bench HardQwen3.6 Plus by 9.134.843.9+9.1
- SciCodeKimi K2.5 by 8.34940.7+8.3
- SWE-bench ProQwen3.6 Plus by 5.950.756.6+5.9
- LMArena · WebDevQwen3.6 Plus by 25 rating points14361461+25 rating
- LiveCodeBench v6Qwen3.6 Plus by 2.18587.1+2.1
- SWE-bench VerifiedQwen3.6 Plus by 276.878.8+2
- SWE-bench Multilingualtie7373.8tie
AgentsQwen3.6 Plus
- MCP AtlasQwen3.6 Plus by 10.363.874.1+10.3
- GDPValQwen3.6 Plus by 7.517.224.7+7.5
- τ-Bench V3 · BankingQwen3.6 Plus by 6.614.220.8+6.6
- OSWorld-Verifiedtie63.362.5tie
ReasoningKimi K2.5
- Humanity's Last ExamKimi K2.5 by 2.930.727.8+2.9
- GPQA Diamondtie87.988.2tie
- CritPttie3.12.9tie
FactsQwen3.6 Plus
- AA-Omniscience · Non-hallucinationQwen3.6 Plus by 31.134.365.4+31.1
- SimpleQA VerifiedQwen3.6 Plus by 9.834.344.1+9.8
- AA-Omniscience · AccuracyKimi K2.5 by 8.835.226.4+8.8
Images and chartsQwen3.6 Plus
- CharXiv (RQ)Qwen3.6 Plus by 477.581.5+4
- Video-MMEKimi K2.5 by 3.287.484.2+3.2
- MMMU-ProQwen3.6 Plus by 2.675.478+2.6
MathQwen3.6 Plus
- IMOAnswerBenchQwen3.6 Plus by 281.883.8+2
- HMMT Feb. 2026tie87.187.8tie
- AIME 2026tie95.895.3tie
Long documentsQwen3.6 Plus
- LongBench v2Qwen3.6 Plus by 16162+1
- AA-LCRtie7878.3tie
Following instructionsQwen3.6 Plus
- IFBenchQwen3.6 Plus by 570.275.2+5
Other results22 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.
- MCPMarkQwen3.6 Plus by 18.729.548.2+18.7
- Tool-DecathlonQwen3.6 Plus by 1227.839.8+12
- ToolathlonQwen3.6 Plus by 1227.839.8+12
- Terminal-Bench 2.0Qwen3.6 Plus by 10.850.861.6+10.8
- AA-OmniscienceQwen3.6 Plus by 8.2-7.30.9+8.2
- Artificial Analysis Coding IndexQwen3.6 Plus by 7.746.854.5+7.7
- AA Agentic IndexQwen3.6 Plus by 7.321.729+7.3
- NL2RepoQwen3.6 Plus by 5.93237.9+5.9
- HMMT Nov. 2025Qwen3.6 Plus by 5.489.294.6+5.4
- WideSearchKimi K2.5 by 4.77974.3+4.7
- τ³-BenchQwen3.6 Plus by 4.76670.7+4.7
- frontiermath_tier_4_v1Qwen3.6 Plus by 4.14.28.3+4.1
- SimpleVQAKimi K2.5 by 3.971.267.3+3.9
- MATH-VisionQwen3.6 Plus by 3.884.288+3.8
- AA IntelligenceQwen3.6 Plus by 3.523.527+3.5
- VideoMMMUKimi K2.5 by 2.686.684+2.6
- OmniDocBench 1.5Qwen3.6 Plus by 2.488.891.2+2.4
- LiveBenchQwen3.6 Plus by 1.869.170.8+1.8
- τ²-Bench Telecom (AA run)Qwen3.6 Plus by 1.895.997.7+1.8
- MMLU-ProQwen3.6 Plus by 1.487.188.5+1.4
- HMMT 2025tie95.494.6tie
- HLE (with tools)tie50.250.6tie
Questions people ask
Which is better, Kimi K2.5 or Qwen3.6 Plus?
Qwen3.6 Plus wins seven of the eight areas we test: coding, agents, facts, images and charts, math, long documents and following instructions. Kimi K2.5 wins reasoning.
Which is better for coding?
Qwen3.6 Plus. It wins 6 of the 8 coding tests both models report; Kimi K2.5 wins 1, and 1 is a tie.
Which is cheaper?
Kimi K2.5 costs $0.60 per million input tokens and $3.00 per million output tokens; Qwen3.6 Plus costs $0.50 and $3.00. That makes Qwen3.6 Plus about 3% cheaper for the same work.
How do you compare the two?
We use the 49 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 22 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.