GLM-5.1 vs Kimi K2.6
Wins 1 of 7 areas
Agents
Wins 4 of 7 areas
Coding · Reasoning · Math · Long documents
Kimi K2.6 wins more areas, narrowly.GLM-5.1 is better at agents.
Scores updated · 67 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 software05Kimi K2.65 of 8 tests · 3 ties
- MathCompetition and research-level math04Kimi K2.64 of 4 tests
- ReasoningHard problems that need careful thinking03Kimi K2.63 of 3 tests
- Long documentsFinding answers in very long texts01Kimi K2.61 of 1 test
- AgentsCarrying out multi-step tasks on its own42GLM-5.14 of 6 tests
- FactsGetting facts right instead of making them up11Even1 each · 1 tie
- Following instructionsDoing exactly what it is asked00Even0 each · 1 tie
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.
Kimi K2.6 costs 14% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Kimi K2.6 pulls ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+20.4points ahead
- Harvard-MIT high-school math contest problemsHMMT Feb. 2026+10.1points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+9.7points ahead
Where GLM-5.1 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+10.6points ahead
- Finds the root cause of IT system incidentsAA IT-Bench SRE+9.1points ahead
- Real work tasks from 44 professionsGDPVal+4points ahead
Every test, side by side
All 67 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingKimi K2.6
- SciCodeKimi K2.6 by 6.744.851.5+6.7
- Terminal-Bench 2.1Kimi K2.6 by 4.161.865.9+4.1
- SWE-bench VerifiedKimi K2.6 by 476.280.2+4
- LiveCodeBench v6Kimi K2.6 by 3.985.789.6+3.9
- SWE-bench MultilingualKimi K2.6 by 3.473.376.7+3.4
- Terminal-Bench Hardtie43.243.9tie
- SWE-bench Protie58.458.6tie
- LMArena · WebDevtie15081509tie
AgentsGLM-5.1
- τ-Bench V3 · BankingKimi K2.6 by 9.713.623.3+9.7
- AA IT-Bench SREGLM-5.1 by 9.140.331.2+9.1
- BrowseCompKimi K2.6 by 779.386.3+7
- GDPValGLM-5.1 by 43127+4
- MCP AtlasGLM-5.1 by 3.771.868.1+3.7
- Terminal-Bench 4.0GLM-5.1 by 1.520.5+1.5
ReasoningKimi K2.6
- Humanity's Last ExamKimi K2.6 by 7.430.137.5+7.4
- GPQA DiamondKimi K2.6 by 4.386.891.1+4.3
- CritPtKimi K2.6 by 3.44.68+3.4
FactsEven
- AA-Omniscience · Non-hallucinationGLM-5.1 by 10.670.159.5+10.6
- AA-Omniscience · AccuracyKimi K2.6 by 8.923.732.6+8.9
- SimpleQA Verifiedtie3434.9tie
MathKimi K2.6
- FrontierMath Tiers 1-3 (v2)Kimi K2.6 by 20.436.857.2+20.4
- HMMT Feb. 2026Kimi K2.6 by 10.182.692.7+10.1
- IMOAnswerBenchKimi K2.6 by 2.283.886+2.2
- AIME 2026Kimi K2.6 by 1.195.396.4+1.1
Other results41 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.
- IOI 2025Kimi K2.6 by 128.5 rating points456.5585+128.5 rating
- Apex (Pass@1)Kimi K2.6 by 12.511.524+12.5
- ProfBench (Search)Kimi K2.6 by 104656+10
- ToolathlonKimi K2.6 by 9.340.750+9.3
- PinchBenchKimi K2.6 by 981.290.2+9
- AgentWorldBench - SWEKimi K2.6 by 6.752.158.8+6.7
- Apex-Shortlist (no tools)Kimi K2.6 by 6.371.177.4+6.3
- Vals.ai Financial Agent 1.1 - without web searchGLM-5.1 by 6.260.254+6.2
- Artificial Analysis Coding IndexKimi K2.6 by 655.861.8+6
- Apex-Shortlist (with tools)GLM-5.1 by 5.87973.2+5.8
- CritPt (no tools)Kimi K2.6 by 5.43.79.1+5.4
- GDPval-AA (Elo)GLM-5.1 by 53 rating points15351482+53 rating
- AgentWorldBench - TerminalKimi K2.6 by 5.247.352.5+5.2
- MCPAtlas Public (Pass@1)GLM-5.1 by 5.271.866.6+5.2
- AgentWorldBench - SearchKimi K2.6 by 522.527.5+5
- GPQA (unspecified)Kimi K2.6 by 4.986.191+4.9
- AA-OmniscienceKimi K2.6 by 4.40.85.3+4.4
- SciCode (subtask)Kimi K2.6 by 4.347.752+4.3
- AA Agentic IndexGLM-5.1 by 3.125.222.1+3.1
- Apex Shortlist (Pass@1)Kimi K2.6 by 3.172.475.5+3.1
- IFBench (prompt loose) (Tools-allowed)GLM-5.1 by 2.976.673.7+2.9
- TauBench V3 - AverageKimi K2.6 by 2.769.772.4+2.7
- IMOAnswerBench (with tools)Kimi K2.6 by 2.691.193.7+2.6
- AgentWorldBench - MCPGLM-5.1 by 2.467.665.2+2.4
- Terminal-Bench 2.0GLM-5.1 by 2.36966.7+2.3
- AgentWorldBench - OverallKimi K2.6 by 2.151.353.4+2.1
- frontiermath_tier_4_v1Kimi K2.6 by 2.112.514.6+2.1
- LiveBenchKimi K2.6 by 270.272.2+2
- Vals.ai Financial Agent 1.1 - with web searchGLM-5.1 by 1.960.758.8+1.9
- τ²-Bench Telecom (AA run)GLM-5.1 by 1.897.795.9+1.8
- AgentWorldBench - OSKimi K2.6 by 1.759.160.8+1.7
- HLE (with tools)Kimi K2.6 by 1.752.354+1.7
- AgentWorldBench - WebGLM-5.1 by 1.351.550.2+1.3
- TauBench V3 - RetailGLM-5.1 by 1.284.182.9+1.2
- MMLU-ProKimi K2.6 by 1.18687.1+1.1
- SWE-QAGLM-5.1 by 1.172.771.6+1.1
- AA Intelligencetie26.127tie
- MMLU-ProX (avg en/de/fr/es/it/ja/zh/hi/pt/ko) (Tools-allowed)tie85.885tie
- TauBench V3 - Airlinetie8585.8tie
- AgentWorldBench - Androidtie59.158.9tie
- Finance Agent v2tie44.844.9tie
Questions people ask
Which is better, GLM-5.1 or Kimi K2.6?
Kimi K2.6 wins four of the seven areas where both have results: coding, reasoning, math and long documents. GLM-5.1 wins agents. They are level on facts and following instructions.
Which is better for coding?
Kimi K2.6. It wins 5 of the 8 coding tests both models report; GLM-5.1 wins none, and 3 are ties.
Which is cheaper?
GLM-5.1 costs $1.38 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 14% cheaper for the same work.
How do you compare the two?
We use the 67 benchmark tests both models have published scores on. The verdict counts the 26 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 41 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.