Benchgen
Models/moonshot-ai/

Kimi-k1.5

DraftPublic
Model CardTraining jobsEvaluations

Model Details

Kimi K1.5

Organization License Released

Quick answer: Kimi K1.5 is Moonshot AI's January 2025 model scoring 88.3% C-Eval and 87.4% MMLU. Proprietary — predecessor to the Kimi K2 series.

At a Glance

Where Kimi K1.5 leads

  • 88.3% C-Eval — excellent Chinese knowledge breadth
  • 87.4% MMLU — strong general knowledge
  • Early 2025: established track record

Where it lags

  • Proprietary: no open weights
  • Superseded by Kimi K2 series (2025)

Best for: Moonshot API Chinese + English knowledge tasks (C-Eval focus); legacy K1 workloads.

What Kimi K1.5 Is

Kimi K1.5 is the January 2025 release from Moonshot AI — the "K1" generation predecessor to the K2 series. The 88.3% C-Eval demonstrates strong Chinese-language coverage, while the 87.4% MMLU shows solid English academic performance.

Specifications

FieldValue
OrganizationMoonshot AI
LicenseProprietary
Release dateJanuary 22, 2025
ModalityText only

Pricing

Available via Moonshot AI (kimi.ai) API.

Public Benchmark Scores

BenchmarkScoreSourceDate
C-Eval88.3%Benchgen evaluation2025-01
MMLU87.4%Benchgen evaluation2025-01

Kimi K1.5 vs Alternatives

ModelC-EvalMMLULicense
Kimi K1.588.3%87.4%Proprietary
Kimi K2.0905 Instruct——Proprietary

Kimi K2 series supersedes K1.5. For current Moonshot: use Kimi K2 or later.

Related Models & Benchmarks

Frequently Asked Questions

What is Kimi K1.5? Moonshot AI's January 2025 model scoring 88.3% C-Eval, 87.4% MMLU. Proprietary — superseded by K2 series.

Specs from Moonshot AI's Kimi K1.5 release (January 2025) and Benchgen evaluations. Last updated 2026-07-24.

Benchmark Leaderboards

This model isn’t on any benchmark leaderboard yet.