AI
OpenAI cuts GPT-5.6 Luna prices 80%, adds fast mode for Sol
OpenAI announced sweeping price reductions across its GPT-5.6 model family on July 30. The sharpest cut hits GPT-5.6 Luna, now $0.20 per million input tokens and $1.20 per million output — an 80% reduction — while GPT-5.6 Terra drops 20% to $2 input and $12 output. The company also introduced Fast mode for GPT-5.6 Sol in the API, delivering up to 2.5x the speed of standard processing at 2x the standard price with no change in intelligence. Sam Altman framed the strategy as offering "the best price/intelligence tradeoff at every level," with OpenAI crediting efficiency gains across the model, inference stack, and agentic harness for making the cuts possible.
The cuts have immediate downstream effects for developers. Auto-review in the ChatGPT app and Codex CLI is being upgraded from GPT-5.4 to GPT-5.6 Luna, and combined with the new pricing, OpenAI expects that workflow to cost roughly 10 times less than before. The lower prices are also reflected in how usage is counted inside Codex and ChatGPT Work, meaning the savings extend beyond the raw API.
Google DeepMind launches Gemini Robotics 2 with fine-motor and multi-robot capabilities
Google DeepMind unveiled Gemini Robotics 2 on July 30, a new suite of models that lets robots reason through individual movements to handle tasks previously out of reach, including tying delicate knots and coordinating multi-robot workflows. DeepMind CEO Demis Hassabis highlighted a demo with Apptronik's Apollo 2 humanoid using whole-body intelligence to pack for a sports game, and a separate test in which Apollo and Duo robots collaborated to tidy a messy garage — with a high-level reasoning model deciding what to do, where to move, and when to hand off control between units.
This briefing is part of the archive.
Get the app for full archive access — every daily briefing, sourced and attributed.