Kimi K3 vs Qwen3.8 Max Preview
Wins 3 of 8 areas
Reasoning · Facts · Long documents
Wins 5 of 8 areas
Coding · Agents · Images and charts · Math · Following instructions
Qwen3.8 Max Preview is the stronger all-rounder.Kimi K3 is better at reasoning.
Both rank among the ten best models we track in agents and long documents.
Scores updated · 43 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.
- AgentsCarrying out multi-step tasks on its own15Qwen3.8 Max Preview5 of 6 tests
- MathCompetition and research-level math03Qwen3.8 Max Preview3 of 3 tests
- CodingWriting and fixing software23Qwen3.8 Max Preview3 of 5 tests
- Images and chartsUnderstanding pictures, charts and video01Qwen3.8 Max Preview1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.8 Max Preview1 of 1 test
- ReasoningHard problems that need careful thinking30Kimi K33 of 4 tests · 1 tie
- FactsGetting facts right instead of making them up21Kimi K32 of 3 tests
- Long documentsFinding answers in very long texts10Kimi K31 of 1 test
Images and charts, 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.
Qwen3.8 Max Preview costs 56% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.8 Max Preview pulls ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+26.3points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+24.4points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+7.3points ahead
Where Kimi K3 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+15.9points ahead
- Programming problems, refreshed regularlyLiveBench · Coding+8.6points ahead
- Reasons across sets of long documentsAA-LCR+8.4points ahead
Every test, side by side
All 43 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.8 Max Preview
- LiveBench · CodingKimi K3 by 8.681.572.9+8.6
- SciCodeKimi K3 by 7.459.552.1+7.4
- Terminal-Bench 2.1Qwen3.8 Max Preview by 3.88588.8+3.8
- LiveBench · Agentic CodingQwen3.8 Max Preview by 2.562.264.7+2.5
- LMArena · WebDevQwen3.8 Max Preview by 17 rating points16551672+17 rating
AgentsQwen3.8 Max Preview
- Terminal-Bench 4.0Qwen3.8 Max Preview by 26.312.638.9+26.3
- AA IT-Bench SREKimi K3 by 7.447.740.3+7.4
- GDPValQwen3.8 Max Preview by 6.651.758.3+6.6
- τ-Bench V3 · BankingQwen3.8 Max Preview by 1.84647.8+1.8
- OSWorld-VerifiedQwen3.8 Max Preview by 1.384.886.1+1.3
- AA ApexAgentsQwen3.8 Max Preview by 1.141.342.4+1.1
ReasoningKimi K3
- CritPtKimi K3 by 5.723.417.7+5.7
- Humanity's Last ExamKimi K3 by 3.846.943.1+3.8
- LiveBench · ReasoningKimi K3 by 2.590.788.2+2.5
- GPQA Diamondtie93.592.8tie
FactsKimi K3
- AA-Omniscience · Non-hallucinationQwen3.8 Max Preview by 24.446.871.2+24.4
- AA-Omniscience · AccuracyKimi K3 by 15.947.631.7+15.9
- SimpleQA VerifiedKimi K3 by 3.350.647.3+3.3
Images and chartsQwen3.8 Max Preview
- MMMU-ProQwen3.8 Max Preview by 2.380.582.8+2.3
MathQwen3.8 Max Preview
- FrontierMath Tier 4Qwen3.8 Max Preview by 7.33946.3+7.3
- LiveBench · MathematicsQwen3.8 Max Preview by 6.984.491.3+6.9
- FrontierMath Tiers 1-3 (v2)Qwen3.8 Max Preview by 2.572.274.7+2.5
Following instructionsQwen3.8 Max Preview
- LiveBench · Instruction FollowingQwen3.8 Max Preview by 2.771.474.1+2.7
Other results19 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.
- Agents' Last ExamQwen3.8 Max Preview by 24.128.352.4+24.1
- DeepSWE 1.1Kimi K3 by 12.46956.6+12.4
- DeepSWEKimi K3 by 10.967.556.6+10.9
- AA-OmniscienceKimi K3 by 7.719.712+7.7
- MLS Bench Litelower is betterQwen3.8 Max Preview by 7.348.341+7.3
- ProgramBench (Almost Solved)Kimi K3 by 717.510.5+7
- livebench_languageKimi K3 by 5.885.579.7+5.8
- GDPval-AA v2Qwen3.8 Max Preview by 57 rating points16821739+57 rating
- Artificial Analysis Coding IndexKimi K3 by 4.476.271.8+4.4
- Toolathlon VerifiedKimi K3 by 476.572.5+4
- HLE (with tools)Kimi K3 by 3.659.856.2+3.6
- NL2RepoKimi K3 by 2.15855.9+2.1
- AA IntelligenceQwen3.8 Max Preview by 1.843.645.4+1.8
- CyberGymKimi K3 by 1.58078.5+1.5
- AA Agentic IndexKimi K3 by 150.649.6+1
- Toolathlontie73.272.5tie
- Agents' Last Exam (ALE-CLI)tie27.627tie
- JobBenchtie52.953.4tie
- livebench_data_analysistie78.778.4tie
Questions people ask
Which is better, Kimi K3 or Qwen3.8 Max Preview?
Qwen3.8 Max Preview wins five of the eight areas we test: coding, agents, images and charts, math and following instructions. Kimi K3 wins reasoning, facts and long documents.
Which is better for coding?
Qwen3.8 Max Preview. It wins 3 of the 5 coding tests both models report; Kimi K3 wins 2.
Which is cheaper?
Kimi K3 costs $3.00 per million input tokens and $15.00 per million output tokens; Qwen3.8 Max Preview costs $2.00 and $6.00. That makes Qwen3.8 Max Preview about 56% cheaper for the same work.
How do you compare the two?
We use the 43 benchmark tests both models have published scores on. The verdict counts the 24 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 19 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.