Model profile
Kimi K2.6
Kimi K2.6 by Moonshot AI scores 74/100 on BenchLM's provisional leaderboard (#20 of 79), with an Arena Elo of 1461 and a 256K-token context window. API pricing is $0.95/$4 per million input/output tokens. Newer replacements: Kimi K2.7 Code.
Evidence coverage
60 of 296 tracked benchmarks are published. 33 are verified and 27 provisional. 7 of 8 categories are measured.
- Published / tracked
- 60 / 296
- Verified
- 33
- Provisional
- 27
- Categories measured
- 7 / 8
Evidence by category
- Agentic22 benchmarksMixed evidence
- Coding13 benchmarksMixed evidence
- Reasoning2 benchmarksReported
- Knowledge10 benchmarksMixed evidence
- Math5 benchmarksVerified
- Multilingual0 benchmarksNot measured
- Multimodal7 benchmarksMixed evidence
- Inst. Following1 benchmarkReported
According to BenchLM.ai, Kimi K2.6 ranks #20 out of 79 models on the provisional leaderboard with an overall score of 74/100. It also ranks #13 out of 35 on the verified leaderboard. This places it in the mid-tier of AI models, with strengths in specific benchmark categories.
Kimi K2.6 is a open weight model with a 256K token context window. It uses explicit chain-of-thought reasoning, which typically improves performance on math and complex reasoning tasks at the cost of higher latency and token usage.
BenchLM links it directly to Kimi K2.5 as the earlier related model in that lineage. This profile currently has 60 of 296 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.
Its strongest category is Mathematics (#2), while its weakest is Multimodal & Grounded (#29). This performance profile makes it particularly strong for mathematical reasoning, scientific computing, and quantitative analysis.
Peer position
Exact provisional scores and ranks for the closest listed peers.
Range 74.0–75.0
- GPT-5.2OpenAICompare#1975.0GPT-5.2 is #19 with a score of 75.0.
- Kimi K2.6Current modelMoonshot AI#2074.0Kimi K2.6 is #20 with a score of 74.0.
- DeepSeek V4 Pro (High)DeepSeekCompare#2174.0DeepSeek V4 Pro (High) is #21 with a score of 74.0.
- GPT-5.5 ProOpenAICompareUnranked75.0GPT-5.5 Pro is Unranked with a score of 75.0.
- Holo3-35B-A3BH CompanyCompareUnranked74.0Holo3-35B-A3B is Unranked with a score of 74.0.
- MiMo-V2.5-ProXiaomiCompareUnranked74.0MiMo-V2.5-Pro is Unranked with a score of 74.0.
- SWE-1.7CognitionCompareUnranked74.0SWE-1.7 is Unranked with a score of 74.0.
Category percentile
More
Relative position among models eligible for each sourced category. A higher percentile means a stronger position within that category's ranked cohort; 100 is highest.
- Math90%Eligible cohort rank #2 of 11Category score 71.8
- Coding95%Eligible cohort rank #6 of 101Category score 88.1
- Agentic86%Eligible cohort rank #18 of 120Category score 80.2
- Multimodal74%Eligible cohort rank #29 of 110Category score 68.0
Category evidence
Scores and ranks appear only where this model has published benchmark evidence. Categories without displayable source records remain not measured.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #18 of 120Percentile 86thWeight 22%22 benchmarksMixed sources | 80.2 | #18 of 120 | 86th | 22% | 22 benchmarks | Mixed sources |
| CodingRank #6 of 101Percentile 95thWeight 20%13 benchmarksMixed sources | 88.1 | #6 of 101 | 95th | 20% | 13 benchmarks | Mixed sources |
| ReasoningRank Not rankedWeight 17%2 benchmarksReported | 0.0 | Not ranked | Not available | 17% | 2 benchmarks | Reported |
| KnowledgeRank Not rankedWeight 12%10 benchmarksMixed sources | 63.4 | Not ranked | Not available | 12% | 10 benchmarks | Mixed sources |
| MathRank #2 of 11Percentile 90thWeight 5%5 benchmarksVerified | 71.8 | #2 of 11 | 90th | 5% | 5 benchmarks | Verified |
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | Not ranked | Not available | 7% | 0 benchmarks | Not measured |
| MultimodalRank #29 of 110Percentile 74thWeight 12%7 benchmarksMixed sources | 68.0 | #29 of 110 | 74th | 12% | 7 benchmarks | Mixed sources |
| Inst. FollowingRank Not rankedWeight 5%1 benchmarkReported | 0.0 | Not ranked | Not available | 5% | 1 benchmark | Reported |
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Chatbot Arena performance
Scroll horizontally to inspect confidence intervals and vote counts.
| View | Elo | Confidence interval | Votes |
|---|---|---|---|
| Text Overall | 1461 | ±4.7 | 32,657 |
| Coding | 1512 | ±7.5 | 8,944 |
| Math | 1485 | ±14.3 | 1,718 |
| Instruction Following | 1457 | ±6.8 | 10,517 |
| Creative Writing | 1432 | ±9.1 | 5,061 |
| Multi-turn | 1458 | ±8.7 | 5,687 |
| Hard Prompts | 1487 | ±5.5 | 21,037 |
| Hard Prompts (English) | 1487 | ±7.0 | 10,286 |
| Longer Query | 1477 | ±6.6 | 13,360 |
Benchmark Details
Rows below have a displayable published verification record. Each source link and provenance note remains in the page HTML while its category is closed. Source-unverified manual rows and generated rows stay hidden.
Agentic22 benchmarks
Artificial Analysis Agentic Index
GDPval-AA normalized
Gert Labs Composite Game Benchmark
Artificial Analysis Briefcase
Artificial Analysis AutomationBench
Artificial Analysis EnterpriseOps-Gym
Artificial Analysis Harvey LAB-AA
Artificial Analysis ITBench-AA
Artificial Analysis Tau3-Banking
Coding13 benchmarks
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
Software Engineering Benchmark Verified
Scientific Code Benchmark
Vibe Code Bench v1.1
Artificial Analysis Coding Index
Artificial Analysis SciCode
Artificial Analysis Terminal-Bench v2.1
Reasoning2 benchmarks
Artificial Analysis Long Context Reasoning
Critical Physics Tasks
Knowledge10 benchmarks
Humanity's Last Exam
Graduate-Level Google-Proof Q&A
GPQA Diamond
Artificial Analysis GPQA Diamond
Artificial Analysis Humanity's Last Exam
Artificial Analysis Omniscience Index
Artificial Analysis Omniscience Accuracy
Artificial Analysis Omniscience Hallucination Rate
Artificial Analysis Openness Index
Math5 benchmarks
FrontierMath v2 Tiers 1-3
AIME 2026
Harvard-MIT Mathematics Tournament February 2026
FrontierMath v2 Tier 4
Multimodal7 benchmarks
Massive Multi-discipline Multimodal Understanding Pro
CharXiv Reasoning
MMMU-Pro with Python
Artificial Analysis MMMU-Pro
Design Arena Website Elo
Inst. Following1 benchmark
Artificial Analysis IFBench
Frequently Asked Questions
How does Kimi K2.6 perform overall in AI benchmarks?
Kimi K2.6 currently ranks #20 out of 79 models on BenchLM's provisional leaderboard with an overall score of 74. It also ranks #13 out of 35 on the verified leaderboard. It is created by Moonshot AI and features a 256K context window.
Is Kimi K2.6 good for knowledge and understanding?
Kimi K2.6 has visible benchmark coverage in knowledge and understanding, but BenchLM does not currently assign it a global category rank there.
Is Kimi K2.6 good for coding and programming?
Kimi K2.6 ranks #6 out of 79 models in coding and programming benchmarks with an average score of 88.1. It is among the top performers in this category.
Is Kimi K2.6 good for mathematics?
Kimi K2.6 ranks #2 out of 79 models in mathematics benchmarks with an average score of 71.8. It is among the top performers in this category.
Is Kimi K2.6 good for reasoning and logic?
Kimi K2.6 has visible benchmark coverage in reasoning and logic, but BenchLM does not currently assign it a global category rank there.
Is Kimi K2.6 good for agentic tool use and computer tasks?
Kimi K2.6 ranks #18 out of 79 models in agentic tool use and computer tasks benchmarks with an average score of 80.2. There are stronger options in this category.
Is Kimi K2.6 good for multimodal and grounded tasks?
Kimi K2.6 ranks #29 out of 79 models in multimodal and grounded tasks benchmarks with an average score of 68. There are stronger options in this category.
Is Kimi K2.6 good for instruction following?
Kimi K2.6 has visible benchmark coverage in instruction following, but BenchLM does not currently assign it a global category rank there.
Is Kimi K2.6 open source?
Yes, Kimi K2.6 is an open weight model created by Moonshot AI, meaning it can be downloaded and run locally or fine-tuned for specific use cases.
Does Kimi K2.6 have full benchmark coverage on BenchLM?
Not yet. Kimi K2.6 currently has 60 published benchmark scores out of the 296 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.
What is the context window size of Kimi K2.6?
Kimi K2.6 has a context window of 256K, which determines how much text it can process in a single interaction.
Related Resources
Don't miss the next GPT moment
Which models moved up, what is new, and what it costs. One email each week.
Free. One email per week.