Models49in catalog
Largest context1.0MFw Kimi K3

Avg context

321K

Cheapest input

Free/M

Kimi K2 Thinking 251104

Largest output

1.0M

Kimi K3

Speed tiers

balanced 43deep 6

Capabilities

What Moonshot models support across the catalog.

Vision

29 / 49

59% of models

Tool use

46 / 49

94% of models

Function calling

46 / 49

94% of models

Extended thinking

25 / 49

51% of models

Streaming

49 / 49

100% of models

Prompt caching

32 / 49

65% of models

All Moonshot models

Sorted by context window, largest first. Tap a row for full specs.

ModelContext
Fw Kimi K31.0M
Kimi K31.0M
Kimi K3 (batch)1.0M
Kimi K3 Fast1.0M
Kimi K3 Us1.0M
Databricks Kimi K31M
In Moonshotai Kimi K31M
Moonshotai Kimi K31M
Fw Kimi K2 5262K
Fw Kimi K2 6262K
Fw Kimi K2 7 Code262K
Kimi K2 0905262K
Kimi K2 0905 Preview262K
Kimi K2.5262K
Kimi K2 5 Fp4262K
Kimi K2.6262K
Kimi K2.6 (free)262K
Kimi K2.7 Code262K
Kimi K2.7 Code (batch)262K
Kimi K2 Instruct 0905262K
Kimi K2p5262K
Kimi K2p6262K
Kimi K2p6 Fast262K
Kimi K2p7 Code262K
Kimi K2p7 Code Fast262K

Showing 25 of 49 models

See rankings

How Moonshot models rank against the full catalog.