Hugging Face 发布 Qwen3.8-27B 智能体适配器
Hugging Face 在 Pi、Hermes Agent 和 OpenCode 三种智能体框架内,为 Qwen3.8-27B 分别训练了 LoRA 适配器,并将其合并为可复用的超级适配器。官方同时公开适配器、GGUF、强化学习任务、训练流程和 safeagents 插件。Hugging Face 称,Pi 适配器将 Terminal-Bench 2.1 的 pass@1 从 39% 提升至 55%,每任务平均 Token 用量从 2.9K 降至 1.5K,但这些基准结果仍属厂商自报。
该项目提供了一条可复现路径,可将开放模型专门适配到特定编程智能体框架,而不必仅依赖更大或更通用的模型。
Combining RL on a custom task distribution with your favorite agent harness yields more accurate and efficient models for that target harness and task distribution.
Models are not fully optimized for all harnesses and the tasks you’re doing. Instead, they target general performance. So when you combine a model and harness, you’re leaving performance and efficiency on the table for the stuff you’re actually doing. With RL, you can train a custom harness adapter that squeezes everything out of your model for your harness and task combination.
Take Qwen3.8-27B, its Terminal-Bench 2.1 score with the pi harness is 39% which uses on average 2.9K tokens per task. If we freeze those weights and add a custom RL-trained LoRA adapter we can get 55% pass@1, with average token usage of 1.5K. Most of this is down to the model reasoning less, calling less tools, and performing better first time.
Our earlier work (multi-harness RL, harness adapters) showed that RL on your own task distribution, inside your own harness, gives a more accurate and more efficient model. Those experiments covered data-analysis and code-building tasks. Here we wanted to push this further by exploring whether the result holds at 27B, on harder real-world software engineering tasks, across three harnesses: Pi, OpenCode and Hermes Agent. And most of all, share a fully reusable recipe, so you can reproduce this on your own task, harness, model combinations.
#Hugging Face#Qwen3.8-27B#TRL#AsyncGRPOTrainer#OpenEnv#Harbor#Pi#Hermes Agent#OpenCode#Terminal-Bench 2.1#SWE-bench Pro hard-51#safeagents