Agentic RL inside the harness: how LFM2.5 trains with sandboxes and tools
helloiamleonie · x · 2026-09-25
Leonie details the agentic RL stage: since models are mostly used through agent harnesses, train them inside the harness so they're already familiar with system prompts and built-in tools. Components: agentic office tasks (presentations, analysis reports), the harness the model works through, and a fresh sandbox per rollout with tool and filesystem access.
More from coding & agent
- The decade of agents: simulate real customers to test your agent before launch — EXM7777 · 2026-11-11
- FluentCRM MCP Server lets you manage marketing automation from Cursor or Claude — modelcontextprotocol · 2026-11-11
- Databricks' revised Genie agent tells CEO Ali Ghodsi the unvarnished truth — TechCrunch AI · 2026-11-11
- Elvis Saravia: agents will subsume traditional software — and can even cut code bloat — omarsar0 · 2026-11-11
- Every launches Everycut, an agent-native Mac video editor where every control is an agent tool — every · 2026-11-11
- Anthropic's Fable 5.2 built a professional video editor overnight, unsupervised — every · 2026-11-11