NVIDIA · Nemotron 3.5
Small open hybrid Mamba-MoE (30B total / 3B active) tuned for fast, high-volume agent tasks and for customization on a single GPU.
Leader: Claude Opus 5.5 at 58
Leader: Claude Opus 5.5 at 66.9%
Leader: Claude Mythos 5.1 at 60.9%
Leader: Claude Fable 5.1 at 91.4%
Leader: Claude Sonnet 5 at 85.2%
Leader: Claude Opus 5.5 at 1846
Leader: GPT-6 Astra at 91.5%
Leader: GPT-6 Astra at 96.1%
Leader: Claude Opus 5.5 at 61.4%
Leader: GPT-5.6 Sol at 32.3%
Leader: Nemotron 3 Ultra 550B A55B at 86.8%
Leader: Kimi K3 at 88.7%
| Result | Score | Reported by | Source |
|---|---|---|---|
| tau2-Bench Banking (AA) · reasoning | 8.9% | independent | Artificial Analysis: Nemotron 3.5 Lightning ↗ |
| AA-Omniscience Index · reasoning | -17.7 index | independent | Artificial Analysis: Nemotron 3.5 Lightning ↗ |
| PinchBench | 85.37% | vendor | NVIDIA Nemotron 3.5 Lightning model card (NVIDIA-measured comparison table) ↗ |
| SWE-bench Multilingual | 39.33% | vendor | NVIDIA Nemotron 3.5 Lightning model card (NVIDIA-measured comparison table) ↗ |
| IFBench · loose | 71.88% | vendor | NVIDIA Nemotron 3.5 Lightning model card (NVIDIA-measured comparison table) ↗ |