Model Terminal
Kimi K2.6
Kimi K2.6 is Moonshot AI's open-weight multimodal agentic foundation model built for long-horizon coding, tool use, and autonomous agent workflows. It supports text, image, and video inputs via a native vision encoder (MoonViT), features a 256K-token context window, and a sparse Mixture-of-Experts architecture with 1.04T total parameters and 32B active per forward pass. Source: supabase.
Value score
Not scored
Context
—
tokens
Max output
—
tokens
Price
—
Capability radar
Peer value bars
Identity
- Developer
- ~moonshotai
- Openness
- —
- Modalities
- —
- Release
- —
- Knowledge cutoff
- —
- Deprecation
- —
- API docs
- —
Benchmarks
No benchmark scores yet.
Compare nearby
Pricing
- Input / 1M
- —
- Output / 1M
- —
- Speed
- —