GPT-6.1 Terra scores 40.7 on Bug Hunt Bench at max, but collapses below that tier
PawelHuryn · x · 2026-10-27
Pawel Huryn tested GPT-6.1 Terra on Bug Hunt Bench (2 real repos, 105 planted bugs):
- GPT-6.1 Terra (max): 40.7 — cost and time effective
- GPT-5.6 Terra (max): 32
- GPT-5.6 Terra (xhigh): 20 / (high): 18
- GPT-6.1 Terra (xhigh): 11.3 / (high): 9.7
Key takeaway: performance drops off a cliff right after the max tier, falling far below Haiku 5.5 (xhigh).
More from coding & agent
- Forcing personal agents to explain tool calls makes opaque background work auditable — _AustinCalvert_ · 2026-10-27
- A single prompt replaces six-figure competitive analysts with weekly AI monitoring briefs — alex_verem · 2026-10-27
- Seroter's reading list: AI coding workflows, subagent tradeoffs, and AI's zero financial payoff — rseroter · 2026-10-27
- A one-page project template with 9 fields for clearer scope and ownership — dSebastien · 2026-10-27
- Marketstack MCP Server brings stock market data APIs to AI assistants — modelcontextprotocol · 2026-10-27
- DMARKOFF MCP connector adds DMARC, SPF & DKIM analytics to AI assistants — modelcontextprotocol · 2026-10-27