NVIDIA Nemotron 3.5 Bolt Card - AA-Omniscience accuracy: leaderboard
NVIDIA Bolt 3.5 release card: AA-Omniscience accuracy. Score (%) from the publisher's own evaluation table; source-specific task, offline and precision conditions remain separate. Full-precision and native 24GB IQ2 runs are separate; MCP-Atlas uses the 500 public tasks.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Rough model projection: around 2026. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma4 26B-A4B-BF16 | 22.17 | |
| 2 | Nemotron 3.5 Bolt 74B-A9B (BF16) | 21.8 | |
| 3 | Qwen3.6-35B-A3B-BF16 | 19.47 | |
| 4 | Nemotron 3.5 Bolt 74B-A9B-24GB (IQ2) | 16.9 | |
| 5 | Nemotron Lightning 3.5 30B A3B (BF16) | 16.63 |
Interactive version: theaggregate.ai/benchmark?slug=nvidia-nemotron-3-5-bolt-card-aa-omniscience-accuracy · How It Works · Data refreshed daily, snapshot 2026-10-16.