HOT RADAR
opensourceHEAT 10.9°

AI CURATED HOT STORY · 11/19/2026

Hugging Face Releases Qwen3.8 Coding LoRA Adapter Boosting Success to 65%

1 reports archived1 independent sourcesupdated 11/19/2026, 00:14:50
Synthesis & Latest Updates
1 Sources Cross-Validated

Hugging Face released Qwen3.8-27B-pi-adapted, a LoRA adapter trained for the pi coding agent that lifts SWE-bench Pro pass@1 from 39.2% to 64.7% and Terminal-Bench 2.1 from 39.5% to 55.8%. The approach pairs one small adapter per model and harness, trained with AsyncGRPO reinforcement learning inside pi and rewarded for fewer tool calls, cutting median turns on solved tasks from 59 to 23 while using roughly 4.5x fewer tokens than the base model's thinking mode. The rank-32 LoRA ships under Apache-2.0 and runs on vLLM or as a GGUF adapter in llama.cpp.

LATEST/Hugging Face released Qwen3.8-27B-pi-adapted, a LoRA adapter trained for the pi coding agent that lifts SWE-bench Pro pass@1 from 39.2% to 64.7% and Terminal-Bench 2.1 from 39.5% to 55.8%. The approach pairs one small adapter per model and harness, trained with AsyncGRPO reinforcement learning inside pi and rewarded for fewer tool calls, cutting median turns on solved tasks from 59 to 23 while using roughly 4.5x fewer tokens than the base model's thinking mode. The rank-32 LoRA ships under Apache-2.0 and runs on vLLM or as a GGUF adapter in llama.cpp.

Hourly heat curve

21 fully observed hours
Now
10.9
Peak
17.405:00
24h change
-37%

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first