Condesion-Bench - Allocation Conditions: leaderboard
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.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 100 |
| 2 | GPT-5 Mini | 100 |
| 3 | O3 (2025-04-16) | 99.6 |
| 4 | O4 Mini (2025-04-16) | 98.01 |
| 5 | GPT-OSS-20B | 91.63 |
| 6 | GPT-OSS-120B | 87.65 |
| 7 | Mistral Large 2 (Jul) | 82.47 |
| 8 | GPT-4.1 Mini | 71.31 |
| 9 | Claude 3.5 Haiku | 71.31 |
| 10 | Claude 3.5 Sonnet (20241022) | 70.92 |
| 11 | Llama 3.3 70B Instruct | 69.72 |
| 12 | GPT-4.1 | 69.72 |
| 13 | Gemini 2.0 Flash (001) | 63.35 |
| 14 | Gemini 2.0 Flash Lite (001) | 45.42 |
| 15 | Mistral Small 3.2 | 45.02 |
Interactive version: theaggregate.ai/benchmark?slug=condesion-bench-allocation-conditions · How It Works · Data refreshed daily, snapshot 2026-10-07.