Dev + SWE-2 High Fusion Mode Inside Devin Called a 'Holy Grail' Combo
CtrlAltDwayne · x · 2026-09-15
A user reports running GPT-6 Astra with SWE-2 High in Fusion mode inside Devin, calling the results 'unreal' and noting the combo lets them stretch their remaining 21% Codex usage. The post is a first-hand excitement share without concrete configs or benchmarks — a lead worth checking against the attached media, not a verified evaluation.
More from coding & agent
- DeepSeek's full-stack code output lags Claude Code and Cursor — devs trade prompting tips — AdEnvironmental4143 · 2026-09-15
- OpenAI Article: Migrating to the Agents API for More Reliable Agents — OpenAIDevs · 2026-09-15
- Scoble on his dumbest agent workflow, and Plasma Radio: Slack for AI agents — Scobleizer · 2026-09-15
- First open-source MCP server for FirstRate5 bridges NatHERS energy ratings and LLMs — Total-Concern6676 · 2026-09-15
- Dev uses ChatGPT Pro to write full tech spec for a Lemon Squeezy alternative — CtrlAltDwayne · 2026-09-15
- GitHub Copilot auto model selection adds efficiency, balance, and intelligence tiers — mariorod1 · 2026-09-15