AI
Mid-Harness: Scaling Actions Between Model and Harness for Terminal Agents
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 29/09/2026, 20:00:00 UTC
Briefing created: 01/10/2026, 08:11:46 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 01/10/2026, 05:18:24 UTC.
Briefings
The study introduces Mid-Harness, a method that samples and verifies candidate actions before execution to improve terminal agent reliability. On TerminalBench-Lite, a GPT-5.6 Sol verifier increased Pass@1 from 50.00% to 68.03% using 8 sampled actions. Pairwise verification with TMAX-9B yielded the best results, and distilling its responses further improved performance. Mid-Harness also enhanced performance across other models and benchmarks, identifying action scaling as a promising target for test-time compute.
Key points
- Mid-Harness samples and verifies candidate actions before forwarding one for execution to improve action reliability.
- On TerminalBench-Lite, a GPT-5.6 Sol verifier raised Pass@1 from 50.00% to 68.03% with 8 sampled actions.
Context and limitations
Results are limited to TerminalBench-Lite and TMAX-9B model; performance improvements observed across other models and harnesses remain unverified.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 01/10/2026, 08:11:46 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeiaqkzlqoklgc25ejsfvezhygno5gqjbgizlzgjl3e5tmy2qlomohy