Thinking Machines just unveiled Inkling 2 Small, a 594B-parameter MoE model with only 17B active parameters built for autonomous agentic coding and visual reasoning. It can run experiments, fix bugs, and refine interfaces using visual feedback in the loop. Full weights are slated to drop under Apache 2.0 next week, but you can already test it on Tinker Playground.
Thinking Machines just launched Inkling, an open-weights MoE model featuring 975B total and 41B active parameters alongside a 1M token context window. Pretrained across text, images, and audio, it lets developers dial thinking effort up or down to balance cost against latency. Why it matters: Open-source builders finally get a frontier-scale multimodal base designed from the ground up for agentic fine-tuning.
Thinking Machines Lab co-founders Barret Zoph and Luke Metz just left Mira Murati's fledgling startup to rejoin OpenAI, alongside early staffer Sam Schoenholz. The sudden move comes amid allegations that Zoph shared confidential data with competitors, though OpenAI dismissed the concerns and put him in a leadership role. Losing key technical founders is a massive blow to Murati's $12B venture in the middle of a brutal talent war.
Tinker just opened to general availability, ditching its waitlist for all users. The AI customization platform also launched fine-tuning support for the trillion-parameter Kimi K2 Thinking reasoning model, added vision input via Qwen3-VL, and introduced an OpenAI-compatible sampling API that lets developers test models even mid-training.
Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, just unveiled Tinker, an automated tool designed to fine-tune frontier open-source models. Developers can use a few lines of code to run supervised fine-tuning or reinforcement learning on Meta's Llama and Alibaba's Qwen, bypassing messy GPU cluster management entirely. Access is open via application for free during beta.
