Remember when AI picked models for you? Those days are over. This week, we shipped GPT-5 on Augment with a model picker—now you choose between Claude Sonnet 4 and GPT-5, run comparisons, and see which handles your codebase better.

OpenAI claims GPT-5 has "PhD-level expertise" with 45% fewer hallucinations (we'll see), GitHub Models launches with multiple GPT-5 variants, and security researchers warn AI coding tools are creating new blind spots. Plus, Google wants your GitHub workflow.

Augment now offers users the choice to directly select GPT-5 alongside Claude Sonnet 4

Augment Code has integrated OpenAI’s new GPT‑5 model into its platform and introduced a model picker, allowing users to choose between Claude Sonnet 4 (still the default) and OpenAI GPT‑5. This rollout has begun and will continue to reach all users soon.

This update is key for developers and teams using Augment —now they can explicitly opt into the capabilities of GPT‑5 for coding tasks. It signals a shift from behind-the-scenes model selection to user-directed control, enabling comparative workflows and performance testing directly within the IDE.

Empower your team: leverage the model picker to compare Claude Sonnet 4 and GPT‑5 on identical tasks —like refactors, test generation, or bug fixes— so you can determine which model aligns better with your project's specific demands and workflow preferences.

The current tests were run with Claude Sonnet 4’s reasoning mode disabled, so evaluation results may change when reasoning is turned on. A follow-up comparison is expected, which could shift conclusions for more complex tasks.

OpenAI launched GPT-5, calling it their "smartest, fastest, most useful model yet" with PhD-level expertise across multiple domains. The model significantly reduces hallucination rates by 45% compared to GPT-4o and introduces unified reasoning capabilities. CEO Sam Altman described the leap from GPT-4 feeling like "talking to a college student" to GPT-5 feeling like "talking to a PhD-level expert". The model is now available to all ChatGPT users, including free tier users, marking the first time OpenAI has provided reasoning capabilities without a subscription.

GitHub made GPT-5 generally available through GitHub Models, offering multiple variants including gpt-5, gpt-5-mini, gpt-5-nano, and gpt-5-chat. The integration provides developers with direct access to OpenAI's latest model through GitHub's development platform, alongside models from Cohere, DeepSeek, Meta, and Microsoft. This represents a significant shift in making cutting-edge AI accessible to the developer community.

Google released Gemini CLI GitHub Actions in beta, providing a no-cost AI coding teammate for repositories. The tool offers intelligent issue triage, accelerated pull request reviews, and on-demand collaboration through @gemini-cli mentions. It includes enterprise-grade security with credential-less authentication and granular control features. The launch represents Google's push into AI-assisted development workflows directly within GitHub.

Microsoft announced GPT-5 integration across consumer, developer, and enterprise products. The unified system is built to understand when to respond quickly versus thinking more deeply, representing a significant advancement in Microsoft's AI strategy. This integration affects millions of developers using Microsoft's development tools and Azure services.

Security researchers warned that AI coding tools like Claude CLI are "unintentionally changing the attack surface for developers and businesses". The report highlights how AI-generated code can introduce new vulnerabilities and create security blind spots in development workflows.

Intro to “Pointer,” a long-range modeling approach aimed at cutting inference costs and data hunger—relevant for scaling context windows efficiently.

Another week of separating AI signal from noise. If we saved you from a demo that would've crashed prod, we've done our job.

📧 Got a story? Reply with your AI tool wins, fails, or war crimes. Best stories get featured (with credit).

📤 Share the skepticism: Forward to an engineer who needs saved from the hype. They'll thank you.

✍️ Who's behind this? The Augment Code team—we build AI agents that ship real code. Started this newsletter because we're tired of the BS too.

🚀 Try Augment: Ready for AI that gets your whole codebase?