Condesion-Bench: 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 decision satisfaction rate in percent, the share of actions that satisfy all of their conditions, over trading days from October 2024 to September 2025; higher is better.
Metric: Decision Satisfaction Rate (%, from a 0-1 share times 100): share of actions that satisfy all their conditions, 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
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 Mini (2025-08-07) | 91.24 | #352 |
| 2 | GPT-5 | 89.24 | #185 |
| 3 | O3 (2025-04-16) | 88.84 | #238 |
| 4 | O4 Mini (2025-04-16) | 86.06 | #375 |
| 5 | GPT-OSS-120B | 75.3 | #712 |
| 6 | GPT-OSS-20B | 70.52 | #1188 |
| 7 | Claude 3.5 Sonnet (20241022) | 63.75 | #651 |
| 8 | Claude 3.5 Haiku | 57.37 | #1597 |
| 9 | GPT-4.1 Mini | 52.99 | #799 |
| 10 | Mistral Large 2 (Jul) | 51.4 | #1164 |
| 11 | GPT-4.1 | 51.39 | #506 |
| 12 | Llama 3.3 70B Instruct | 46.22 | #1394 |
| 13 | Mistral Small 3.2 | 33.07 | #1647 |
| 14 | Gemini 2.0 Flash (001) | 32.27 | #733 |
| 15 | Gemini 2.0 Flash Lite (001) | 28.69 | #1122 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=condesion-bench · How It Works · Data refreshed daily, snapshot 2026-11-22.