Skip to content
Model Pareto

Kimi K2.7 Code

Moonshot AI (Kimi)Open weightsartificialanalysis.ai ↗
Compare with nearest rivalsCompare with…
Input
$0.95/M tok
Output
$4.00/M tok
Blended
$1.712/M tok · 3:1
Speed
66tok/s
Context
256Ktokens

Long-tail import from Artificial Analysis (2026-09-24); AA variant: Kimi K2.7 Code. Price = AA median host price. Card: "adopts the same native int4 quantization method as Kimi-K2-Thinking".

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.

Hardware
16× H200 141GB
FP8 weights
Throughput
~26,263 tok/s
many requests batched
Price
$0.28–2.34 /M tok
busy → light use
On your own machine
Too large
needs data-center GPUs
Assumptions (5)
  • FP8 weights (1000B params, 32B active per token) + 25% KV-cache headroom ≈ 1250 GB
  • 16× H200 141GB at $3.59–$7.91/GPU-hour on-demand (2026-09-24)
  • ~26,263 output tok/s aggregate at batch 1111 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~198,704 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Coding

SWE-bench VerifiedEstimated
~#8 of 14 (estimated)~76.8%
  1. #1Claude Opus 596.0%
  2. #2Claude Sonnet 585.2%
  3. #6Mistral Medium 3.577.6%
  4. ~#8Kimi K2.7 CodeEstimated76.8%
  5. #8Muse Glimmer76.0%
No published score — estimated from related benchmarks
SWE-bench Pro (Public)Estimated
~#14 of 19 (estimated)~56.0%
  1. #1Claude Opus 5.589.9%
  2. #2Claude Sonnet 5.581.3%
  3. #13Hy357.9%
  4. ~#14Kimi K2.7 CodeEstimated56.0%
  5. #14Inkling54.3%
No published score — estimated from related benchmarks
DeepSWE v1.1Estimated
~#27 of 31 (estimated)~53.9%
  1. #1DeepSeek V4.1 Flash74.2%
  2. #2Grok 4.772.6%
  3. #26Nex-N2.5-Pro55.8%
  4. ~#27Kimi K2.7 CodeEstimated53.9%
  5. #27Qwen3.8 Max (0902)51.0%
No published score — estimated from related benchmarks
LiveCodeBenchEstimated
~#8 of 20 (estimated)~77.6%
  1. #1Qwen3.8-Omni-Flash92.6%
  2. #2Qwen3.8 Flash-Next91.9%
  3. #7gpt-oss-20b77.7%
  4. ~#8Kimi K2.7 CodeEstimated77.6%
  5. #8Mistral Small 464.0%
No published score — estimated from related benchmarks
AA Coding Agent IndexEstimated
~last of 15 (estimated)below measured range (<41.1)
  1. #1Claude Opus 5.566.0
  2. #2Claude Fable 5.162.2
  3. #14GPT-6 Luna41.1
  4. ~#15Kimi K2.7 CodeEstimated<41.1
No published score — estimated from related benchmarks

Writing & chat

MMMLUEstimated
~#3 of 5 (estimated)~86.1%
  1. #1Gemini 3.1 Pro92.6%
  2. #2Gemma 4 31B88.4%
  3. ~#3Kimi K2.7 CodeEstimated86.1%
  4. #3Claude Haiku 4.583.0%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Estimated
~#36 of 54 (estimated)~1405 Elo
  1. #1Claude Opus 5.51501
  2. #2Claude Fable 5.1Few votes1496
  3. #35Nemotron 3 Ultra1405
  4. ~#36Kimi K2.7 CodeEstimated1405
  5. #36Mistral Large 31397
No published score — estimated from related benchmarks
LMArena Text (Elo)Estimated
~#36 of 54 (estimated)~1420 Elo
  1. #1Claude Opus 5.51509
  2. #2Claude Fable 5.11501
  3. #35Muse GlimmerFew votes1424
  4. ~#36Kimi K2.7 CodeEstimated1420
  5. #36Qwen3.5 122B A10B1417
No published score — estimated from related benchmarks

Hard reasoning

AIME (latest)Estimated
~#4 of 25 (estimated)~94.3%
  1. #1Inkling97.1%
  2. #2Muse Glimmer94.7%
  3. #3Nova 2.0 Lite94.3%
  4. ~#4Kimi K2.7 CodeEstimated94.3%
  5. #4gpt-oss-120b93.4%
No published score — estimated from related benchmarks

Agents

OSWorld-VerifiedEstimated
~#10 of 12 (estimated)~69.5%
  1. #1Kimi K384.8%
  2. #2Qwen3.8 27B84.3%
  3. #9Nex-N2.5-Mini71.2%
  4. ~#10Kimi K2.7 CodeEstimated69.5%
  5. #10Muse Glimmer65.9%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~last of 5 (estimated)below measured range (<47.5%)
  1. #1GLM-5.3 Flash59.1%
  2. #2Kimi K358.3%
  3. #4GPT-5.547.5%
  4. ~#5Kimi K2.7 CodeEstimated<47.5%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#15 of 18 (estimated)~68.6%
  1. #1Atria Dawn Preview92.5%
  2. #2GPT-6 Astra91.5%
  3. #14Inkling77.1%
  4. ~#15Kimi K2.7 CodeEstimated68.6%
  5. #15Mistral Medium 3.548.6%
No published score — estimated from related benchmarks