Benchmarks/Kimi K2.6
Moonshot AI / Open weights

Kimi K2.6

An open-weight model with competitive intelligence and agentic evidence.

INTELLIGENCE54index
CODINGNOT MEASURED
AGENTIC1520Elo
SPEEDNOT MEASURED
FIT / INTERPRETATION

Where the evidence points.

  • Open deployment
  • Tool use
  • Cost-sensitive agents
TECHNICAL CONTEXT

Score needs conditions.

Context
Not disclosed
Input / 1M
Not recorded
Output / 1M
Not recorded
Access
Open weights
EVIDENCE TRAIL

Original measurements.

Kimi K2.6 evaluationOPEN SOURCE ↗