Bespoke Labs post-trains Inkling with SFT+GRPO, gaining 57pp on repo-level coding and 40% token efficiency
AlexGDimakis · x · 2026-09-04
Bespoke Labs researchers post-trained the Inkling base model on a specific GitHub repo using SFT on strong-teacher trajectories plus GRPO RL in repo-specialized environments. SFT alone added 52pp on the held-out fontTools eval; RL lifted it to 57pp, with good transfer to Terminal-Bench 2.1 and SWE-Bench Lite and 40% better token efficiency.
More from coding & agent
- MCP Platform NitroStack Wins Product Hunt Product of the Day, Focuses on Retention — EstablishmentFun4373 · 2026-10-21
- CodeRabbit Ships 5,000 PRs a Week With a Tiny Team, Hires by Strategy-Game Hours — omojumiller · 2026-10-21
- LukeW maps the five UI problems every AI-native company faces with agents — LukeW · 2026-10-21
- Claude Code over ADB guts a bloatware-laden Android TV, freeing ~90% more RAM — edwinarbus · 2026-10-21
- Developer uses Claude to fix his adware-slowed Xiaomi TV, releases tool for any Android TV — edwinarbus · 2026-10-21
- Open-weights Ornith-1.5-122B debuts with 84.5% SWE-bench Verified, runs on a single DGX card — _akhaliq · 2026-10-21