Moonshot AI · Model release
Kimi K2.6: open-weight volume with a smaller context window
Kimi K2.6 trades frontier context for open weights and low API rates — a sensible volume and self-hosting candidate when 256K context is enough.
What changed
- Moonshot AI released K2.6 as a 1T-parameter MoE model with open weights.
- API pricing sits well below closed frontier tiers.
- Context is limited to 256K tokens relative to K3 and the closed leaders.
Best fit
- Cost-sensitive volume
- Self-hosting
- Avoiding vendor lock-in
What to be careful about
- 256K context is a hard ceiling for long-repository or document work.
- MoE serving still needs serious hardware despite active-parameter efficiency.
Dataset snapshot: Kimi K2.6 is also represented in the checked comparison dataset. Prices and benchmark figures carry their own field dates.