Competing at Every Price Point with Agentic Evolution over a Menu of LLMs
2026-08-17 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors created a system called RoboPhD that uses a mix of different language models (LLMs) to do tasks better and cheaper than competitors. By evolving programs step-by-step and starting with only a small set of examples, RoboPhD can generate code and find scientific papers more accurately or at lower cost than others. It beats almost all competitors on two different task leaderboards by always offering the best trade-off between price and accuracy. This shows it is possible to build agents that outperform others at every price level using careful evolution across available models.
LLM (Large Language Model)Pareto dominationagentic evolutioncode generationscientific document retrievalmeta-agentexecution-checked codetask-specific optimizationprice-performance tradeoffbenchmark datasets
Authors
Andrew Borthwick
Abstract
Consider a firm that surveys its competition for a particular agentic task and seeks to offer superior accuracy at every competitor price point. A firm that Pareto-dominated its competitors would leave no rational customer a reason to buy elsewhere. This paper shows a path to this kind of capability via agentic evolution over a menu of LLMs, from training pools of at most 100 examples. Given a priced menu of nine LLM endpoints; brief documentation of the task, objective, and API; a simple seed agent; and an operator-chosen per-problem cost target - usually set at an incumbent's own price - RoboPhD, an evolutionary meta-agent, evolves complete agent programs that attack the public frontiers of two semantically dissimilar tasks point by point: DS-1000 (execution-checked code generation) and PaperFindingBench (LLM-judged scientific document retrieval). Our officially scored submissions hold every Pareto-frontier slot but one on the two tasks' leaderboards, including Pareto domination of both the top-scoring and the lowest-cost competing points.