KIMI K2

Open Source

Moonshot AI•Released on 2026-01-20

Moonshot AI's open-source flagship with top HLE and Live Codebench scores. HN users praise its agentic coding ability approaching Claude Haiku 4.5, making it the coding king among open-source models.

77
Overall Score

Core Specs

128K
Context Window
32K
Max Output
✓
Reasoning
✓
Open Source
Multimodal Support
text

Scenario Scores

Coding
#7 / 1488
Writing
#13 / 1470
Translation
#12 / 1475
Data Analysis
#11 / 1478
Conversation
#13 / 1472
Image Gen
—

User Feedback Highlights

Based on community feedback. Hover to see original reviews.

+ Good tool calling+ Open source & free+ Strong coding ability− Slightly weaker English− Smaller ecosystem+ Top HLE score− Incomplete documentation
Sentiment:👍 72%😐 18%👎 10%

Pros & Cons

Pros

  • +Open source & free
  • +Strong coding ability
  • +Top HLE score
  • +Good tool calling

Cons

  • −Smaller ecosystem
  • −Slightly weaker English
  • −Incomplete documentation

Reliability

SLA99.0%
Incidents (30d)0
支持自托管。API 服务器主要在中国。
View Status Page →

Pricing

Input (per 1M tokens)$0.20
Output (per 1M tokens)$0.60
Free trial available
Updated on 2026-03-06

Compare with Others

KIMI K2 vs GPT-5.4 →KIMI K2 vs GPT-5.4 Thinking →KIMI K2 vs Claude Opus 4.6 →

Benchmarks

sweBench71.3%
hle85%
liveCodebench82%