Condesion-Bench - Contextual Conditions: leaderboard
Metric: Satisfaction rate (%) of contextual conditions (feasibility tied to the scenario context), 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 Mini | 97.21 |
| 2 | GPT-5 | 96.81 |
| 3 | O4 Mini (2025-04-16) | 96.41 |
| 4 | O3 (2025-04-16) | 96.41 |
| 5 | Claude 3.5 Sonnet (20241022) | 93.63 |
| 6 | GPT-OSS-120B | 86.06 |
| 7 | Claude 3.5 Haiku | 85.66 |
| 8 | GPT-OSS-20B | 84.46 |
| 9 | GPT-4.1 | 79.68 |
| 10 | Gemini 2.0 Flash (001) | 77.29 |
| 11 | GPT-4.1 Mini | 76.1 |
| 12 | Llama 3.3 70B Instruct | 72.51 |
| 13 | Mistral Large 2 (Jul) | 71.71 |
| 14 | Mistral Small 3.2 | 66.53 |
| 15 | Gemini 2.0 Flash Lite (001) | 60.16 |
Interactive version: theaggregate.ai/benchmark?slug=condesion-bench-contextual-conditions · How It Works · Data refreshed daily, snapshot 2026-10-07.