BenchLeader
Menu
LeaderboardsModelsProvidersBenchmarksCompareReleasesBlogMethodology
Moonshot AIOpen weights: moonshotai/Kimi-K2.7-CodeReasoning modelAuto-detected

Kimi K2.7 Code

Best configuration ranks #159 of 610 on the BenchLeader Index at 56.2 ±4.4. Last measured 8 Sept 2026. Released 12 Jun 2026.

Blended price
$1.71/M
$0.950 in · $4.00 out
Output speed
63 tok/s
First answer
38 s
first token 1.20 s
Context
262k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index56
  2. Reasoning60
  3. Coding54
  4. Agents & tools60
  5. Maths55
  6. Knowledge53
  7. Instruction following62
  8. Long context66
  9. Composite47

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

BenchmarkdefaultSourceTrend
GPQA Diamond87.9%#53Epoch AI Benchmarking Hub
SimpleBench57.9%#34SimpleBench
LiveBench Reasoningnot in index82.8%#34LiveBench
GPQA Diamond (AA)not in index89.6%#71Artificial Analysis
Humanity's Last Exam (AA)not in index35.0%#81Artificial Analysis

Coding

BenchmarkdefaultSourceTrend
SciCode47.5%#73SciCode
WeirdML54.1%#58WeirdML
FrontierCode30.1%#18Cognition
LMArena WebDev1472#47LMArena
LiveBench Codingnot in index74.0%#40LiveBench
SciCode (AA)not in index47.8%#81Artificial Analysis
LiveCodeBench82.0%#56Vals AI
SWE-bench (Vals)not in index78.2%#30Vals AI

Agents & tools

BenchmarkdefaultSourceTrend
APEX-Agents27.6%#28Mercor
LiveBench Agentic Codingnot in index45.7%#40LiveBench
Terminal-Bench Hard44.7%#34Artificial Analysis
τ²-Bench Telecom (AA)not in index90.1%#66Artificial Analysis
Terminal-Bench 2.1 (Vals)67.0%#28Vals AI

Maths

BenchmarkdefaultSourceTrend
FrontierMath Tiers 1–354.0%#37Epoch AI Benchmarking Hub
FrontierMath Tier 412.2%#47Epoch AI Benchmarking Hub
OTIS Mock AIME95.6%#29Epoch AI Benchmarking Hub
LiveBench Mathematicsnot in index79.6%#47LiveBench

Knowledge

BenchmarkdefaultSourceTrend
SimpleQA Verified36.5%#40Epoch AI Benchmarking Hub
LiveBench Data Analysisnot in index62.7%#48LiveBench
AA-Omniscience-10.2#140Artificial Analysis

Instruction following

BenchmarkdefaultSourceTrend
LiveBench Languagenot in index77.9%#30LiveBench
IFBench63.1%#114Artificial Analysis

Long context

BenchmarkdefaultSourceTrend
AA-LCR79.3%#65Artificial Analysis

Composite

BenchmarkdefaultSourceTrend
Epoch Capabilities Indexnot in index150.1#44Epoch AI Benchmarking Hub
LiveBench68.4%#45LiveBench
AA Intelligence Index26.3#109Artificial Analysis

Where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

ProviderSpeedFirst tokenInput $/MOutput $/MContextQuantisation
ModelRun [by Modular]143 tok/s0.33 s$0.850$3.75262kfp4
Fireworks80 tok/s0.55 s$0.950$4.00262k–
SiliconFlow70 tok/s0.84 s$0.859$3.80262kfp8
StreamLake63 tok/s1.12 s$0.713$3.00256k–
Cloudflare57 tok/s1.24 s$0.950$4.00262k–
CoreWeave55 tok/s1.78 s$0.710$3.50262kint4
DeepInfra54 tok/s0.92 s$0.680$3.40262kfp4
Inceptron53 tok/s1.20 s$0.706$3.21262kint4
NovitaAI45 tok/s1.46 s$0.912$3.84262kint4
Venice42 tok/s1.54 s$0.750$3.50256kint4
Moonshot AI Highspeed42 tok/s1.83 s$1.90$8.00262kint4
Moonshot AI41 tok/s1.02 s$0.950$4.00262kint4
Baseten40 tok/s0.53 s$0.950$4.00262kfp4
GMICloud39 tok/s2.96 s$0.950$4.00262kfp8
Alibaba Cloud Int.34 tok/s2.54 s$0.950$4.00262kfp8

Price history

Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.

$0.00$0.963$1.93$2.89$3.85Jun 26Jul 26Aug 26
input outputnow $0.706 in · $3.21 out

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.190 per 1M. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.0016$0.001443.0 s
Summarise a 30-page report12,000 / 600$0.014$0.007047.8 s
Code edit6,000 / 1,500$0.012$0.00831.0 min
Agentic coding session60,000 / 4,000$0.073$0.0391.7 min
Structured extraction2,000 / 200$0.0027$0.001641.4 s

See also

Best under $2/MBest open weightsBest for codingBest for agentsAll Moonshot AI models

Data as of 9 Sept 2026. Compare with another model.