TSCG: leaderboard
Metric: Scenario D, 20 tools, json-text baseline: tool-selection accuracy (%) (self-reported). Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 14B | 90.2 |
| 2 | Gemma 3 12B | 85 |
| 3 | Llama 3.1 8B | 78.4 |
| 4 | Gemma 3 4B | 49.9 |
| 5 | Qwen 3 4B | 44.3 |
| 6 | Mistral 7B | 35 |
| 7 | phi-4-14B | 0 |
Interactive version: theaggregate.ai/benchmark?slug=tscg · How It Works · Data refreshed daily, snapshot 2026-10-09.