$ diff kimi-k3-1-pro k2-horizon
Kimi K3.1 Pro vs K2 Horizon
Values come from Kimi K3.1 Pro and K2 Horizon, where each is cited to its source. This page states no benchmark result and ranks neither model — it puts two published specifications next to each other. Where a lab has not published a figure, the row says so rather than guessing.
Moonshot AIKimi K3.1 Pro
- context
- 1.0M
- weights
- —
- $/M in
- $2
- $/M out
- $10
- context
- 524K
- weights
- open
- $/M in
- —
- $/M out
- —
What actually differs
- Context window
- Kimi K3.1 Pro takes 1.0M against 524K — 2× more room in a single request.
- Recency
- Kimi K3.1 Pro shipped 50 days after K2 Horizon (2026-10-23 vs 2026-09-03).
Full spec
| Attribute | Kimi K3.1 Pro | K2 Horizon |
|---|---|---|
| Developer | Moonshot AI | Ifm |
| Released | 2026-10-23 | 2026-09-03 |
| Context window | 1,048,576 tokens ([source](../../sources/blogs/moonshot-2026-10-24-k3-pricing.md)) | 524,288 (flagship 375B-A23B, native from midtraining onward) |
| Pricing | $2.00/MTok input · $10.00/MTok output · $0.20/MTok cached input; cache writes $2.00/MTok (5m), $4.00/MTok (1h); excludes taxes ([source](../../sources/blogs/moonshot-2026-10-24-k3-pricing.md)) | not recorded |
| License | not recorded | Apache 2.0 (weights and code) |
| Availability | Kimi.ai, mobile apps, Kimi Work, Kimi Code, Kimi API; enterprise access off by default at launch ([source](../../sources/blogs/moonshot-2026-10-23-kimi-k3-1-pro.md)) | Hugging Face (weights, code, checkpoints, training data where redistributable) |