NVIDIA Nemotron 49B Report - BFCL V2 Live: leaderboard
BFCL V2 Live: tool use through function calling on user-contributed queries from the Berkeley Function Calling Leaderboard; the model card gives different mode labels for this benchmark. From Table 4 of NVIDIA's Llama-Nemotron technical report, comparing the final LN-Super (Llama-3.3-Nemotron-Super-49B-v1) with DeepSeek-R1-Distill-Llama-70B, QwQ-32B and Llama-3.3-70B-Instruct, with LN-Super shown in both reasoning modes. The report states that all evaluations use a 32k context length, temperature 0.6 with top-p 0.95 for reasoning on and greedy decoding for reasoning off, with up to 16 completions per prompt averaged as pass@1.
Metric: Score (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 Nemotron Super 49B (v1) (Non-reasoning) | 73.9 | #1336 (Llama 3.3 Nemotron Super 49B (v1)) |
| 2 | Llama 3.3 Nemotron Super 49B (v1) (Reasoning) | 73.7 | #1336 (Llama 3.3 Nemotron Super 49B (v1)) |
| 3 | QwQ-32B | 71.6 | #909 |
| 4 | DeepSeek R1 Distill Llama 70B | 65.5 | #1641 |
| 5 | Llama 3.3 70B Instruct | 60.4 | #1353 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=nvidia-nemotron-49b-report-bfcl-v2-live · How It Works · Data refreshed daily, snapshot 2026-11-12.