Distilling GLM-5.2 Traces into Qwen3.5-4B via OpenCode Harness
NielsRogge · x · 2026-08-10
A developer shared a weekend project demonstrating how to distill model capabilities into a smaller agent. By extracting traces from GLM-5.2 using the OpenCode harness, they fine-tuned a Qwen3.5-4B model via Q-LoRA SFT.
Because the model is trained directly against the harness, it effectively learns to utilize OpenCode's built-in tools. The project currently relies solely on SFT without RL, and the resulting model is deployed on Hugging Face Inference Endpoints.
Related event: Fine-Tuning AI Coding Agents with OpenCode and TRL(2 posts)→
More from coding & agent
- DIRECT: retrieval-based next-question suggestions for conversational agents — burkov · 2026-10-16
- Claude Code and Codex now autocomplete the user's next action — tokenbender · 2026-10-16
- Hermes Agent ships official Reverse Engineer Anything plugin with one-command setup — Teknium · 2026-10-16
- MLExpert: 35B MoE Agent Trained to Autonomously Handle Machine Learning Engineering — Shangheng Du · 2026-10-16
- Cowork-Gym: Alibaba Synthesizes 64.9K Agent Work Environments, Boosting Qwen by 8.7 Points — Anni Zou · 2026-10-16
- MMSearch-OPD: Joint RL and On-Policy Distillation for Long-Horizon Multimodal Search Agents — PekingUniversity · 2026-10-16