Auto. Claude by Anthropic. OpenAI's Codex. Which will you choose? You can now pick your preferred model when starting a new agent session on GitHub Mobile. Optimize for speed on the go or power for deep work. The choice is yours. Download the app. 📱 https://github.com/mobile
Already hands-on with Claude Max Pro 20x daily — Opus 4.6 for deep architectural thinking, Sonnet 4.6 for the speed work 😄 But let’s be real, the model is only as good as the engineer steering it. Knowing what good architecture looks like, asking the right questions, and constantly challenging the model’s reasoning is what separates good from great. I do it all the time — push back, stress test the logic, validate assumptions. The tool amplifies your thinking, it doesn’t replace it.
Giving developers the power to choose between speed and deep work, GitHub Mobile just made coding way more flexible and fun!
This is a powerful shift. GitHub The real question is not which model is better, but which model fits the intent of the task. Speed vs depth. Exploration vs precision. Agentic workflows are becoming less about “having AI” and more about architecting AI choice intentionally. Giving developers control at session level is a subtle but strategic move. The future isn’t one model to rule them all — it’s intelligent orchestration.
Which model asks the best follow up questions? I don't need "confidence" I need "cooperation".
the 'choose your model' paradigm is where enterprise AI adoption is heading. teams have different use cases — some need speed, others need depth. letting developers pick per session rather than locking to one provider per org is the right call. from our experience building at Context Studios, we actually split by task type: fast models for tool calls and validation, frontier models for architecture and complex reasoning. any plans to bring model selection to workflow automation beyond mobile?
Well, I think the answer is: it depends. Normally, I use an x1 model for planning and day-to-day tasks. For more complex work, I go with Opus 4.6 it’s expensive, but worth it. I still need to try 5.3 Codex. For the x1 model, I’m currently using Sonnet 4.6.
AI models evolve almost daily, so it’s hard to keep up. The safest approach is usually to work with the latest stable version available.
Claude Opus 4.6 is optimized for high-level thinking and architecture, while Codex 5.3 excels at complex code writing.
Me: I'll just use Auto 😎 Auto: picks the most expensive model 😭