Bug Hunt Bench Adds Per-Model Bug Coverage View; Hunyuan Sits on the Price-Performance Frontier
PawelHuryn · x · 2026-08-28
Pawel Hyrun added a Coverage view to Bug Hunt Bench—a blind-graded board testing frontier coding models against 105 real planted bugs—showing which bugs each model fixes, with filters and PNG export.
He earlier assessed Tencent Hunyuan as "not a breakthrough, but yet another solid partner for everyday coding": putting aside its high Luna scores, it sits on the price-performance Pareto frontier with open weights. The board also offers score-vs-cost and score-vs-time views, with roughly a 200x cost spread across runs.
Related event: Tencent Hunyuan Hy4 Fixes 17 Bugs for $3.13 in Real-Repo Blind Test(4 posts)→
More from coding & agent
- Running a multi-model UI design arena: Paseo orchestrates four design drafts per model — vista8 · 2026-10-16
- Cortex A/B-tests coding-agent skills and rules by replaying developer corrections before keeping them — Samuele Giampieri · 2026-10-16
- Tool calls are the hidden bottleneck killing AI coding speed — Crabbox and 5 fixes — steipete · 2026-10-16
- Agentic sandbox WorldBox doubles GPT-6 Astra's MazeBench score, 23 to 61 gems — ZhitingHu · 2026-10-16
- Two prompts: Gemini 4 Argon turns a single-player game into full multiplayer — chongdashu · 2026-10-16
- Anthropic and Gray Swan launch $300,000 indirect prompt injection red-teaming contest — bcherny · 2026-10-16