GPT-6 Astra Coming to Devin: Benchmark and Cost Claims
Melvin Vivas · X post · 2026-09-04 · Open on X
Topics: Industry Trends & Job Market, LLM Fundamentals, AI Dev Tools & Productivity · Level: intermediate
Summary
Shares a quoted announcement that the GPT-6 Astra model is coming to the Devin coding agent. The announcement says Astra scores within 0.4 points of Fable 5 on FrontierCode 1.1 at 64% lower cost, and sets a new best score on Devin's internal testing benchmark. It's a useful data point on the cost-vs-quality tradeoff when picking models for coding agents.
Key points
- GPT-6 Astra is being added to Devin, an AI coding agent.
- On FrontierCode 1.1, Astra reportedly scores within 0.4 points of Fable 5.
- Astra is said to be 64% cheaper than Fable 5 for about the same benchmark score.
- It is claimed to set a new state-of-the-art (SOTA) on Devin's internal testing benchmark: more thorough tests, clearer reports, better video evidence.
- Takeaway: when choosing a model, compare cost along with benchmark scores. A model that is nearly as good but much cheaper is often the better pick for production agents.
Resources mentioned
- Devin · tool · x.com · paid
A desktop app from the makers of the Devin coding agent for planning, delegating, reviewing and shipping work across fleets of local and cloud coding agents.
Also in: Devin's open-source vulnerability-fixes agent plugin (Melvin Vivas on X · notes), Gemini 4 Argon now available in Devin (Melvin Vivas on X · notes), Devin Defaults to Opus 5.5 in Normal and Fusion Modes (Cloud) (Melvin Vivas on X · notes), Devin: Preview Websites in Its Built-in Cloud Browser (Melvin Vivas on X · notes) and 57 more - GPT-6 Astra · tool · openai.com · paid
The model announced in the quoted launch post, pitched as the developer's most capable model for work, coding, science and cybersecurity, and able to operate a computer.
Also in: Glimpse: A Visual Reasoning Model That Uses a Visual Scratch Space (Melvin Vivas on X · notes), Is AI Getting Cheaper? How to Measure LLM Cost Per Task and Per Intelligence Level (Melvin Vivas on X · notes), GitHub Copilot Adds Intelligent Local Model Routing (Project HydraFusion) (Melvin Vivas on X · notes), Comparing Opus 5.5, GPT-6 Astra and Grok 4.7 on OpenRouter (Melvin Vivas on X · notes) and 55 more - FrontierCode 1.1 · other · cognition.com · free
A coding benchmark used to compare models on software tasks.
Also in: Claude Haiku 5.5 in Devin: beats Sonnet 5 at about 1/8 the cost (Melvin Vivas on X · notes), Claude Opus 5.5 in Devin: #1 on FrontierCode 1.1 (Melvin Vivas on X · notes)
More in Industry Trends & Job Market
- GPT-6 Astra Announced: Computer Use, Coding and Math
- GPT-6 Astra Access Across Plans vs Fable on Claude Max
- GPT-6 Astra Demo: One Agent Juggling Many Computer Tasks
- NVIDIA and Hugging Face announcement on open models
- Mercury 2.5 runs at 1,100 tokens/sec
- Muse Voice Transcribe: Meta's real-time speech-to-text model