Market-Bench - Net Profit Margin: leaderboard

Market-Bench (2026) places LLM-driven retailer agents together in one simulated supply-chain market where they bid for stock in budget-constrained auctions, set prices and write slogans for persona-driven buyers; values are means over 10 runs of a 6-step simulation at temperature 0. This board is cumulative profit over cumulative revenue (at most 1, negative for a loss); higher is better.

Metric: Net profit margin (cumulative profit over cumulative revenue, at most 1, negative for a loss); all 20 models compete as retailer agents in one simulated supply-chain market (sealed-bid procurement auctions, retail pricing and slogans; 6 steps, temperature 0), mean over 10 runs; higher is better. Source: arxiv.org. Saturation forecast: Around April 2027. 19 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 2.5 Flash0.19#463
2Gemini 2.5 Pro0.167#276
3Claude Sonnet 4.50.158#261
4GPT-4o0.117#751
5Phi-40.11#1931
6O30.097#237
7QwenLong-L1-32B0.085#1881
8DeepSeek V3.20.072#419
9Qwen 2.5 VL 32B Instruct0.069#1062
10Market-Bench Llama 3.1 70B (checkpoint unspecified)0.068
11Qwen 2.5 VL 72B Instruct0.058#841
12Market-Bench Qwen2.5 32B (checkpoint unspecified)0.053
13Market-Bench Gemma 3 27B (checkpoint unspecified)0.041
14ERNIE 4.5 300B A47B0.037#1122
15Market-Bench InternLM2.5 20B (checkpoint unspecified)0.031

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=market-bench-net-profit-margin · How It Works · Data refreshed daily, snapshot 2026-11-21.