Nexus Function Calling - Heldout Combined: leaderboard
Metric: Function Call Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 3 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | NexusRaven-V2-13B | 59.9 |
| 2 | GPT-4 Preview (1106) | 48.14 |
| 3 | GPT-3.5 Turbo (1106) | 44.95 |
Interactive version: theaggregate.ai/benchmark?slug=nexus-function-calling-heldout-combined · How It Works · Data refreshed daily, snapshot 2026-10-09.