Noometry

Moonshot AI, open weights

Kimi K2.6

Kimi K2.6 by Moonshot AI ranks 60th of 354 ranked models on the Noometry Index as of October 2026, with a score of 47.7. Its strongest category is writing & preference, where it ranks 26th. API pricing starts at $0.95 per million input tokens and $4 per million output tokens, with a 262K-token context window.

Last verified

Noometry rank
#60 of 354
Index score
47.7
Evidence
Confirmed 51 results
Provider
Moonshot AI
Released
April 20, 2026
Weights
Open weights
Reasoning
Yes
Context window
262K
Max output
262K
Input price
$0.95 / M
Output price
$4 / M
Blended price
$1.71 / M
Output speed
Not measured
Value
#138 of 219
Knowledge cutoff
January 2025
Input
text, image, video
Hugging Face
moonshotai/Kimi-K2.6

Category scores

Each category score combines every public result we have in that category.

  1. Coding 50.7
  2. Agentic & Tool Use 21.9
  3. Reasoning 40.5
  4. Math 57.0
  5. Knowledge 54.0
  6. Multimodal 31.6
  7. Multilingual 54.9
  8. Instruction Following 76.3
  9. Long Context 44.9
  10. Writing & Preference 68.5
CategoryScoreRankResults
Coding50.7#435
Agentic & Tool Use21.9#1374
Reasoning40.5#558
Math57.0#416
Knowledge54.0#544
Multimodal31.6#1033
Multilingual54.9#371
Instruction Following76.3#431
Long Context44.9#521
Writing & Preference68.5#265

Strengths and weaknesses

Categories where Kimi K2.6 places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Writing & Preference68.5+14.8#26 of 312, top 9%
Multilingual54.9+7.5#37 of 297, top 13%
Math57.0+20.5#41 of 327, top 13%

Weakest categories

CategoryScorevs medianRank
Agentic & Tool Use21.9−8.5#137 of 154, top 89%
Multimodal31.6−6.9#103 of 128, top 81%
Long Context44.9+4.0#52 of 296, top 18%

Closest competitors

The models ranked just above and below Kimi K2.6. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
Grok 4#5648.1—1Compare
Kimi K2.5#5748.1$0.9066Compare
Step 5 Preview#5847.9$1.43—Compare
GLM-5.1#5947.8$2.15—Compare
o3#6147.5$3.503Compare
Qwen3.6 Plus#6247.5$1.13—Compare
Inkling-Small#6346.5$0.64—Compare
GPT-5 Pro#6446.4$41.255Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
SWE-bench Verified76.7%#10 of 32, top 32%Epoch AI2026-05-08
LMArena WebDev1509#45 of 113, top 40%LMArena2026-10-08
SciCode53.5%#32 of 121, top 27%Epoch AI
WeirdML55.9%#38 of 119, top 32%Epoch AI
LMArena Coding1488#32 of 294, top 11%LMArena2026-10-08
ALE-Bench1,093#38 of 105, top 37%Epoch AI

Agentic & Tool Use

BenchmarkScorePositionSettingSourceDate
OSWorld 2.04.6%#7 of 9, top 78%Epoch AI
ExploitBench18.4%#6 of 9, top 67%Epoch AI
GBAEval0.9%#17 of 23, top 74%Epoch AI
GDP.pdf12%#32 of 36, top 89%Epoch AI
Vending-Bench 26,205#18 of 60, top 30%Epoch AI

Reasoning

BenchmarkScorePositionSettingSourceDate
NYT Connections (extended)87.2%#26 of 91, top 29%Lech Mazur benchmarks
CritPt8%#54 of 134, top 41%Epoch AI
Chess Puzzles26%#40 of 129, top 32%Epoch AI2026-05-07
EBR-Bench2.4%#24 of 24, top 100%Epoch AI2026-06-25
LMArena Hard Prompts1470#38 of 297, top 13%LMArena2026-10-08
Mystery Game Puzzles18%#47 of 74, top 64%Epoch AI2026-07-17
Mystery Game Puzzles12%noneEpoch AI2026-08-29
DTBench90.9%#34 of 151, top 23%Epoch AI
LMCA37.3%#57 of 125, top 46%Epoch AI
Epoch Capabilities Index151.05#49 of 213, top 24%Epoch AI2026-04-20

Math

BenchmarkScorePositionSettingSourceDate
FrontierMath (Tiers 1-3)57.2%#40 of 81, top 50%Epoch AI2026-06-10
FrontierMath Tier 425.6%#37 of 63, top 59%Epoch AI2026-06-10
MathArena Final-Answer Competitions72.9%#11 of 29, top 38%thinkMathArena
OTIS Mock AIME 2024-202596.1%#30 of 173, top 18%Epoch AI2026-05-02
ProofBench16%#54 of 77, top 71%Epoch AI
LMArena Math1475#32 of 285, top 12%LMArena2026-10-08
FrontierMath (Feb 2025 set)39%#11 of 68, top 17%Epoch AI2026-05-07
FrontierMath Tier 4 (v1)14.6%#16 of 55, top 30%Epoch AI2026-05-08

Knowledge

BenchmarkScorePositionSettingSourceDate
GPQA Diamond90.8%#32 of 186, top 18%Epoch AI2026-05-01
SimpleQA Verified34.9%#48 of 77, top 63%Epoch AI2026-08-10
Vectara Hallucination Rate (lower is better)10.8%#63 of 96, top 66%Vectara Hallucination Leaderboard
LMArena Expert1491#30 of 273, top 11%LMArena2026-10-08

Multimodal

BenchmarkScorePositionSettingSourceDate
LMArena Vision1283#26 of 122, top 22%LMArena2026-10-09
Blueprint-Bench 23.9%#26 of 31, top 84%Epoch AI
Furniture Assembly21.7%#29 of 31, top 94%Epoch AI2026-09-11
LMArena Document1451#22 of 38, top 58%LMArena2026-09-13

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1446#37 of 297, top 13%LMArena2026-10-08
LMArena Chinese1521#25 of 285, top 9%LMArena2026-10-08
LMArena French1471#37 of 223, top 17%LMArena2026-10-08
LMArena German1450#44 of 231, top 20%LMArena2026-10-08
LMArena Japanese1443#28 of 211, top 14%LMArena2026-10-08
LMArena Korean1427#30 of 213, top 15%LMArena2026-10-08
LMArena Russian1446#49 of 283, top 18%LMArena2026-10-08
LMArena Spanish1464#32 of 226, top 15%LMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
LMArena Instruction Following1451#41 of 298, top 14%LMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1468#34 of 291, top 12%LMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1455#38 of 297, top 13%LMArena2026-10-08
LMArena Creative Writing1434#47 of 295, top 16%LMArena2026-10-08
EQ-Bench Creative Writing1725#28 of 115, top 25%EQ-Bench
EQ-Bench 41202#17 of 28, top 61%EQ-Bench
LMArena Multi-Turn1453#49 of 295, top 17%LMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
azure$0.95$4—2026-10-10
deepinfra$0.75$3.50$0.152026-10-10
moonshot$0.95$4$0.162026-10-10
openrouter$0.44$2.45$0.112026-10-10

All Moonshot AI API prices → Estimate your cost →

Compare Kimi K2.6

Other Moonshot AI models

Frequently asked questions

How good is Kimi K2.6?

Kimi K2.6 by Moonshot AI ranks 60th of 354 ranked models on the Noometry Index as of October 2026, with a score of 47.7. Its strongest category is writing & preference, where it ranks 26th. API pricing starts at $0.95 per million input tokens and $4 per million output tokens, with a 262K-token context window.

How much does Kimi K2.6 cost?

Kimi K2.6 costs $0.95 per million input tokens and $4 per million output tokens on Moonshot AI's own API, with cached input at $0.16.

What is Kimi K2.6's context window?

Kimi K2.6 accepts up to 262K tokens of input and can write up to 262K tokens in one response.

Is Kimi K2.6 open source?

Yes. Kimi K2.6's weights are downloadable from Hugging Face (moonshotai/Kimi-K2.6); check the license for commercial terms.

What are Kimi K2.6's strengths and weaknesses?

Relative to other ranked models, Kimi K2.6 places best in writing & preference, multilingual, math and lowest in agentic & tool use, multimodal, long context.

What is Kimi K2.6 best at?

Its best category is writing & preference, where it ranks 26th on Noometry.