Inkling: Thinking Machines Lab's first open-weights model
Inkling is Thinking Machines Lab's Apache-2.0 licensed Mixture-of-Experts model with 975B total and 41B active parameters, trained on 45 trillion multimodal tokens.
What it is
An open-weights multimodal Mixture-of-Experts transformer model released by Mira Murati's Thinking Machines Lab.
What it does
It processes text, images, audio, and video after training on 45 trillion tokens, and is released under an Apache-2.0 license alongside a promised smaller Inkling-Small variant of 276B total and 12B active parameters.
Why it matters
An Apache-2.0 licensed model of this scale from a new lab gives engineers another self-hostable alternative to closed frontier APIs for multimodal workloads, competing with DeepSeek V4 Pro and Kimi K3.
How to use it
Download the weights from Thinking Machines Lab once published and serve them with a compatible inference stack; Inkling-Small is not yet available pending further testing.
Kimi K3: 2.8 Trillion Parameters, Weights Coming July 27
Moonshot AI announced Kimi K3, a 2.8 trillion parameter model available now via API, with open weights promised by July 27, 2026.
Claude Code v2.1.206: /commit-push-pr honors configured push remote and worktree confirmation
Claude Code 2.1.206 makes /commit-push-pr auto-allow pushes to a repo's configured pushDefault remote, not just origin, and adds a /doctor check for trimming CLAUDE.md files.
Claude Code v2.1.207: auto mode ships without opt-in on Bedrock, Vertex AI, and Foundry
Claude Code 2.1.207 turns on auto mode by default for Bedrock, Vertex AI, and Foundry users, removing the prior environment-variable opt-in requirement.