Hey Devs! This week brought faster AI models, better coding agents, new open models, and official Linux support for ChatGPT and Codex. Here are the biggest AI updates for developers from August 8–14, 2026.
Gemini 3.7 Flash becomes Google’s new coding and agent workhorse
Google launched Gemini 3.7 Flash with improvements in coding, web development, tool use, and long-running agent workflows. It is available through the Gemini API, Google AI Studio, and other Google developer tools.
OpenAI makes frontier reasoning up to 14× faster
OpenAI previewed Ultrafast, a new API service tier that runs GPT-5.6 Sol up to 14 times faster and generates up to 750 output tokens per second. Access is currently limited to selected customers.
The official ChatGPT and Codex desktop experience arrives on Linux
OpenAI released a Linux preview of the ChatGPT desktop app with access to projects, local files, and Codex. It supports Ubuntu, Debian, and Fedora on x64 and ARM64 systems.
Read the Linux installation guide
Qwen releases its first Max-class open model
Qwen released Qwen3.8-2.4T-A95B, its first Max-class model with downloadable weights. It is designed for coding, research, professional work, and long-running agent tasks.
Explore the model on Hugging Face
Agent Plugins 1.0 brings portable capabilities to GitHub Copilot
GitHub added support for Agent Plugins 1.0 across VS Code, Copilot CLI, the Copilot SDK, and the Copilot app. Developers can package skills and MCP servers once and reuse them across compatible agent clients.
Meta launches a local model for always-on agents
Meta released Muse Glimmer, a 30-billion-parameter open model designed for local agents, coding, tool use, and long-running workflows. It can run on a Mac or PC with a single consumer GPU.
NVIDIA combines a fast execution model with intelligent model routing
NVIDIA introduced Nemotron 3.5 Lightning for high-volume agent tasks and NeMo Switchyard for routing work across multiple models. Together, they help developers improve the speed, cost, and control of production AI agents.
That’s all for this week! Follow the newsletter to stay updated on the latest AI models, developer tools, and agent workflows. See you in the next issue!