Condesion-Bench - Allocation Conditions: leaderboard

Condesion-Bench (2026) evaluates conditional decision-making in a compositional action space: for one trading day an LLM allocates quantities across 15 S&P 500 stocks under variable, contextual and allocation conditions, and its single zero-shot action is checked against the conditions. This board is the satisfaction rate in percent of allocation conditions, which limit quantities and budget; higher is better.

Metric: Satisfaction rate (%) of allocation conditions (limits on quantities and budget), of the single action an LLM generates zero-shot (temperature 0, JSON output) on the 251 Condesion-Bench instances (one per trading day, October 2024 to September 2025: allocate a budget over 15 S&P 500 stocks under variable, contextual and allocation conditions); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 16 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5100#178
2GPT-5 Mini (2025-08-07)100#345
3O3 (2025-04-16)99.6#231
4O4 Mini (2025-04-16)98.01#370
5GPT-OSS-20B91.63#1183
6GPT-OSS-120B87.65#704
7Mistral Large 2 (Jul)82.47#1144
8GPT-4.1 Mini71.31#789
9Claude 3.5 Haiku71.31#1580
10Claude 3.5 Sonnet (20241022)70.92#648
11Llama 3.3 70B Instruct69.72#1385
12GPT-4.169.72#505
13Gemini 2.0 Flash (001)63.35#727
14Gemini 2.0 Flash Lite (001)45.42#1118
15Mistral Small 3.245.02#1641

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

Interactive version: theaggregate.ai/benchmark?slug=condesion-bench-allocation-conditions · How It Works · Data refreshed daily, snapshot 2026-11-21.