Evaluating 36 open-weight models on SWE-bench Verified, the authors find small models now handle easy and medium tasks well and fail mainly on hard ones. Trying a small model first and escalating to a large one only when needed cut total GPU-hour costs by...
VQV Signal
Research: timeout cascades cut coding-agent GPU costs by up to 33%
Evaluating 36 open-weight models on SWE-bench Verified, the authors find small models now handle easy and medium tasks well and fail mainly on hard ones. Trying a small model first and escalating to a large one only when needed cut total GPU-hour costs by...
Evaluating 36 open-weight models on SWE-bench Verified, the authors find small models now handle easy and medium tasks well and fail mainly on hard ones. Trying a small model first and escalating to a large one only when needed cut total GPU-hour costs by up to 33% without losing accuracy, nearly matching an...
VQV organizes public signals from inspectable sources. It does not independently verify the underlying report.
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
Public Interest components
VQV surfaced this signal because it is recent, relevant to AI Coding Tools, connected to Scobleizer News Pulse (X).
No login, cookies, social SDKs, or automatic posting.