Codex’s harness is fully open source and can run open-weights models
Teknium · x · 2026-07-27
Teknium says Codex is a fully open-source harness: you can audit exactly what happens to your prompt and how it works, and you can run open-weights models through it too.
The post is a direct reply to a complaint about not being able to inspect how “compaction” works, framing Codex as something transparent enough to audit rather than a black box.
Related event: OpenAI Codex Orchestration Layer Goes Fully Open Source(2 posts)→
More from coding & agent
- Team-level AI agents: where should shared context and history live? — Al_Grigor · 2026-09-11
- Chaining dependent MCP tool calls: no rollback, duplicate risk — agentrsdg · 2026-09-11
- DeepMind-led paper makes design docs the source of truth, code disposable — SMART regenerates in 1.5-3h for ~$100 — Roger_M_Taylor · 2026-09-11
- Agent-built classifier labels 192k docs for $0.70 vs $13-26 with frontier LLMs — vanstriendaniel · 2026-09-11
- MathModelAgent gains traction: auto-solves math modeling and writes a submission-ready paper — jihe520 · 2026-09-11
- alphaXiv open-sources OpenResearch to run parallel research agents with any model — alphaXiv · 2026-09-11