TACTL-183: leaderboard
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 | 91.8 |
| 2 | Llama 3.1 405B | 88.5 |
| 3 | DeepSeek R1 Distill Llama 70B | 86.9 |
| 4 | DeepSeek V3 | 86.3 |
| 5 | GPT-4o | 85.2 |
| 6 | Qwen 2.5 72B Instruct | 84.2 |
| 7 | Llama-3.1-Tulu-3-70B | 81.4 |
| 8 | Llama 3.3 70B | 78.7 |
| 9 | Mixtral 8x22B | 65 |
Interactive version: theaggregate.ai/benchmark?slug=tactl-183 · How It Works · Data refreshed daily, snapshot 2026-10-09.