Moonshot AI reportedly readies Kimi K3.1 with 1M-token context, IT Home reports
Model: Kimi K3 · availability, license and releases
Reported by IT Home · not yet confirmed by the company or a second independent outlet. We update this page when it is.
Moonshot AI's API platform exposed a 'kimi-k3-1' model identifier and Kimi's open platform showed a K3.1 preview with a 1M-token context window, IT Home reported on Sept. 28, 2026. The model is expected to offer Low, High, and Max reasoning tiers and may add agent, swarm, search, and batch modes.
China context
- Original name
- 月之暗面
- Outside China
- Not stated in the sources yet
- Claims
- Company-reported; not yet independently evaluated
- For builders
- Developers outside China should monitor Moonshot AI's official API documentation for K3.1 availability and any changes to model identifiers or reasoning parameters.
- For investors
- Investors can track whether Moonshot AI officially launches K3.1 and how its pricing and context length compare with competing long-context models.
Translated from Chinese. Quotes and facts link to the original sources.
On Sept. 28, 2026, IT Home reported that developers found a 'kimi-k3-1' identifier in Moonshot AI's API backend model registry, which passed interface probing and could be called. Later that day, Kimi's official open platform displayed a K3.1 or K3 11111 model preview supporting a 1M-token context window. The report says Kimi K3.1 is expected to support up to 1 million tokens of context and provide Low, High, and Max reasoning intensity options, possibly introducing agent mode, swarm multi-agent collaboration, and task modes such as search and batch processing.
The leaked identifier and platform preview indicate a model with a 1M-token context window and three reasoning intensity tiers (Low, High, Max). The evidence does not include benchmark results, parameter counts, or independent evaluations. Verification question: does the released model match the preview's context length and reasoning tiers?
Developers using Moonshot AI's API may need to adapt to new model identifiers and reasoning-tier options when K3.1 becomes available. The reported 1M-token context and multi-tier reasoning could change API cost and latency trade-offs for long-context applications.
If released as previewed, K3.1 could give Moonshot AI a competitive option for long-context and agent-based workloads, but no pricing or performance data is available in the evidence.
The next observable signal is an official Moonshot AI announcement or API documentation update confirming K3.1 availability, pricing, and supported modes. Until then, the preview remains unconfirmed.