ResearchOctober 02, 2026
-43
🧊
FrugalEvo Cuts AI Code Evolution to $0.55
FrugalEvo solves circle packing for $0.55, roughly 90x cheaper than multi-agent baselines.
#LLM-Guided Evolution#AlphaEvolve#Cost Optimization#Program Synthesis#GPT-5

🔥 What happened
Researchers from NUS and Stanford dropped FrugalEvo — a framework where a pricey LLM plans the strategy and a cheap LLM writes and refines the code. On circle packing it hits new state-of-the-art for $1.68 (GPT-5.6) and just $0.55 (GLM-5.3).
💡 Why it matters
Prior evolution methods like AlphaEvolve optimize per iteration, not per dollar. FrugalEvo beats multi-agent baselines such as CORAL and SwarmResearch — which burn roughly $50 per run — and wins on 9 of 10 tasks under a new Budget-Aware AUC metric.
⚡ Our take
If you're still throwing expensive agent swarms at optimization problems in 2026, you're lighting money on fire. The winning pattern is two-tier: smart model thinks, cheap model grinds.
The title, summary and analysis of this item were produced automatically by an AI system and have not been editorially reviewed. They may contain errors, bias or omissions — when in doubt, read the linked original source.