Newest papers

How language model agents reshape price competition

Martin Bichler, Philip Hierhager, Kassian Köck, Fabian Pieroth · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.21711 · cited by 0 works (OpenAlex) · open-access PDF

Indonesian

Kompetisi Bertrand berulang adalah benchmark yang ketat untuk penetapan harga algoritma: di pasar barang homogen, prosedur pembelajaran standar biasanya memilih hasil permainan Nash dengan fokus harga dekat dengan biaya marginal. Kami menunjukkan bahwa agen penetapan harga LLM dapat melarikan diri dari benchmark ini, tergantung pada keputusan penetapan. LLM melampaui siswa online standar melalui adaptasi yang lebih cepat, tetapi keuntungan ini sebagian besar menghilang terhadap siswa informasi penuh. Algoritma surrogasi yang diekstrak menunjukkan bagaimana LLM dalam penetapan spesifik mendapatkan keuntungan dari secara teratur merendahkan siswa yang lebih lambat. Agen yang memposting satu harga setiap periode dari riwayat pasar mentah yang mencakup harga saingan dapat mempertahankan rezim supra-kompetitif dalam self-play. Agen yang secara berkala mengeluarkan distribusi probabilitas dari resume historis keuntungan mereka sendiri untuk

Machine translation (Meta M2M100); the English original follows.

Abstract

Repeated Bertrand competition is a stringent benchmark for algorithmic pricing: in homogeneous-goods markets, standard learning procedures typically select the stage-game Nash outcome with prices near marginal cost. We show that LLM pricing agents can escape this benchmark, depending on deployment decisions. LLMs outperform standard online learners through faster adaptation, but this advantage largely disappears against full-information learners. An extracted surrogate algorithm illustrates how LLMs in specific deployments profit from regularly undercutting slower learners. Agents that post one price every period from raw market histories that include rival prices can sustain supracompetitive regimes in self-play. Agents that periodically emit probability distributions from summarized histories of their own profits instead move toward the competitive region in self-play. Markets mixing these interfaces can generate dynamic regimes in which LLMs establish a high-price anchor and regularly undercut it to capture demand. These regimes can keep both competitors' prices above marginal cost. Deployment choices shape these outcomes, shifting the focus from stationary equilibria to dynamic pricing regimes.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Economics, Microeconomics, Outcome (game theory), Competition (biology), Profit (economics), Dynamic pricing

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange