DigestAI news desk

Cut through the AI noise.

Agents & Tools22 min read

Qwen3.8-27B-pi fixes coding agent effort levels with RL fine-tuning

Qwen3.8-27B-pi is a fine-tuned version of the Qwen3.8-27B model designed for agentic coding tasks within the Pi agent harness. The model addresses a key flaw in the base model: its effort levels (low, medium, xhigh) were not consistently ordered, with low often spending more reasoning tokens than medium while achieving lower pass rates. The fine-tuning process involved two stages: supervised…

1 source primary source

Key points

  • Qwen3.8-27B-pi fixes inconsistent effort-level reasoning in coding agents via RL fine-tuning, ensuring lower levels spend less reasoning than higher ones
  • On Terminal-Bench 2.1, Pi at medium effort matches base xhigh performance with 41% fewer output tokens and higher pass rates across all benchmarks
  • Open-source release includes BF16/FP8 weights (~30.4 GB) and 17 GGUF variants (9–29 GB), with no official deployment support beyond local use

The results show significant improvements. On Terminal-Bench 2.1, Pi at medium effort matches the base model’s xhigh performance with 41% fewer output tokens. On GPQA Diamond, Pi achieves a pass rate of 86.4%, outperforming the base model across all effort levels. The model is released as open-source under the MIT license, with BF16 and FP8 weights (~30.4 GB) and 17 GGUF variants (9–29 GB). The author, Mario Zechner, notes this is a proof-of-concept release constrained by budget and calls for further exploration, such as scaling RL or testing smaller models through the same pipeline.

Model page: Qwen3.8-27B-pi →

The story so far

2 episodes →
  1. Qwen3.8-27B-pi fixes coding agent effort levels with RL fine-tuningthis story
Full story from huggingface.co · via Reddit AI communities primary sourceOpen source ↗

Qwen3.8-27B-pi: Effort-Ordered Reasoning for Agentic Coding

huggingface.co · 30 September 2026

Loading the full article…

This text was published by huggingface.co. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗

Coverage and discussion

1source
Topics · follow one to build your own front page

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.

Comments

via GitHub Discussions

More in Agents & Tools

All →

Related stories