Moonshot Kimi K3.1 model ID leaks, supports 1M-token context: what are the key facts?

Multiple developers spotted the kimi-k3-1 model identifier in Moonshot's API backend, and it was callable.

Moonshot Kimi K3.1 model ID leaks, supports 1M-token context: what are the key facts?

Kimi's official open platform subsequently showed a K3.1 or K3 11111 model teaser.

Moonshot Kimi K3.1 model ID leaks, supports 1M-token context: what are the key facts?

The model is expected to support a context window of up to 1 million tokens.

Moonshot Kimi K3.1 model ID leaks, supports 1M-token context: what are the key facts?

K3.1 offers three reasoning-effort tiers: Low, High, and Max.

Moonshot Kimi K3.1 model ID leaks, supports 1M-token context: what are the key facts?

The open platform shows K3.1's API pricing is the same as the existing K3.

Physix Frontier · News Briefing Card (enbrief · Sep 29, 2026)

Moonshot Kimi K3.1 model ID leaks, supports 1M-token context

KEY FACTS

Multiple developers spotted the kimi-k3-1 model identifier in Moonshot's API backend, and it was callable.
Kimi's official open platform subsequently showed a K3.1 or K3 11111 model teaser.
The model is expected to support a context window of up to 1 million tokens.
K3.1 offers three reasoning-effort tiers: Low, High, and Max.
The open platform shows K3.1's API pricing is the same as the existing K3.

KEY DATA

1M tokensContext window

3 (Low/High/Max)Reasoning effort tiers

PHYSIX OBSERVATION

By pushing context to the million-token level and introducing multiple reasoning-effort tiers, K3.1 shows Moonshot is shifting the competitive focus from parameters to inference cost and agent capabilities. If pricing holds level with K3, that's good news for developers, but the placeholder may mean final pricing is still subject to change. If multi-agent collaboration ships, it will deal a direct blow to existing agent toolchains.

Source: enbrief original report ↗