GLM-5.2 vs Kimi K2.6
Wins 4 of 7 areas
Coding · Agents · Reasoning · Math
Wins 2 of 7 areas
Long documents · Following instructions
GLM-5.2 wins more areas, narrowly.Kimi K2.6 is cheaper and better at following instructions.
Scores updated · 44 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.
- CodingWriting and fixing software60GLM-5.26 of 7 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own50GLM-5.25 of 6 tests · 1 tie
- MathCompetition and research-level math50GLM-5.25 of 6 tests · 1 tie
- ReasoningHard problems that need careful thinking21GLM-5.22 of 4 tests · 1 tie
- Following instructionsDoing exactly what it is asked02Kimi K2.62 of 2 tests
- Long documentsFinding answers in very long texts01Kimi K2.61 of 1 test
- FactsGetting facts right instead of making them up11Even1 each · 1 tie
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 15% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GLM-5.2 pulls ahead
- Real work tasks from 44 professionsGDPVal+16.7points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+14.2points ahead
- Research-level physics problemsCritPt+12.9points ahead
Where Kimi K2.6 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+8.3points ahead
- Reasons across sets of long documentsAA-LCR+2.7points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+2.7points ahead
Every test, side by side
All 44 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGLM-5.2
- Terminal-Bench 2.1GLM-5.2 by 1277.965.9+12
- LMArena · WebDevGLM-5.2 by 94 rating points16031509+94 rating
- Terminal-Bench HardGLM-5.2 by 6.950.843.9+6.9
- LiveBench · Agentic CodingGLM-5.2 by 4.951.846.9+4.9
- SWE-bench ProGLM-5.2 by 3.562.158.6+3.5
- LiveBench · CodingGLM-5.2 by 1.179.778.6+1.1
- SciCodetie51.251.5tie
AgentsGLM-5.2
- GDPValGLM-5.2 by 16.743.727+16.7
- AA IT-Bench SREGLM-5.2 by 11.542.731.2+11.5
- τ-Bench V3 · BankingGLM-5.2 by 11.334.623.3+11.3
- MCP AtlasGLM-5.2 by 8.776.868.1+8.7
- AA ApexAgentsGLM-5.2 by 5.233.728.5+5.2
- Terminal-Bench 4.0tie10.5tie
ReasoningGLM-5.2
- CritPtGLM-5.2 by 12.920.98+12.9
- Humanity's Last ExamGLM-5.2 by 3.641.137.5+3.6
- GPQA DiamondKimi K2.6 by 1.689.591.1+1.6
- LiveBench · Reasoningtie78.679.4tie
FactsEven
- AA-Omniscience · Non-hallucinationGLM-5.2 by 14.273.759.5+14.2
- AA-Omniscience · AccuracyKimi K2.6 by 8.324.332.6+8.3
- SimpleQA Verifiedtie34.234.9tie
MathGLM-5.2
- LiveBench · MathematicsGLM-5.2 by 5.589.884.3+5.5
- IMOAnswerBenchGLM-5.2 by 59186+5
- FrontierMath Tier 4GLM-5.2 by 3.729.325.6+3.7
- AIME 2026GLM-5.2 by 2.899.296.4+2.8
- FrontierMath Tiers 1-3 (v2)GLM-5.2 by 259.257.2+2
- HMMT Feb. 2026tie92.592.7tie
Following instructionsKimi K2.6
- IFBenchKimi K2.6 by 2.773.376+2.7
- LiveBench · Instruction FollowingKimi K2.6 by 2.162.364.4+2.1
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.
- GDPval-AA v2GLM-5.2 by 318 rating points15081190+318 rating
- HiL-Bench (Tools-allowed)GLM-5.2 by 2543.718.7+25
- AA Agentic IndexGLM-5.2 by 17.339.422.1+17.3
- livebench_data_analysisGLM-5.2 by 8.673.765.1+8.6
- Artificial Analysis Coding IndexGLM-5.2 by 768.861.8+7
- AA IntelligenceGLM-5.2 by 6.733.727+6.7
- τ³-Bench BankingGLM-5.2 by 6.226.820.6+6.2
- SWEBench Pro PublicGLM-5.2 by 3.562.158.6+3.5
- τ²-Bench Telecom (AA run)GLM-5.2 by 3.299.195.9+3.2
- ToolathlonKimi K2.6 by 1.848.250+1.8
- StrongREJECTKimi K2.6 by 1.398.599.8+1.3
- livebench_languageGLM-5.2 by 1.176.275.1+1.1
- AA-Omnisciencetie4.45.3tie
- Global-MMLU-Litetie89.288.4tie
- HLE (with tools)tie54.754tie
Questions people ask
Which is better, GLM-5.2 or Kimi K2.6?
GLM-5.2 wins four of the seven areas where both have results: coding, agents, reasoning and math. Kimi K2.6 wins long documents and following instructions, and costs 15% less. They are level on facts.
Which is better for coding?
GLM-5.2. It wins 6 of the 7 coding tests both models report; Kimi K2.6 wins none, and 1 is a tie.
Which is cheaper?
GLM-5.2 costs $1.40 per million input tokens and $4.40 per million output tokens; Kimi K2.6 costs $0.95 and $4.00. That makes Kimi K2.6 about 15% cheaper for the same work.
How do you compare the two?
We use the 44 benchmark tests both models have published scores on. The verdict counts the 29 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.