Nemotron 3.5 Embed 8B vs Nemotron Colembed VL 8B v2
No benchmarks in this slice.
Specs
| Spec | Nemotron 3.5 Embed 8B | Nemotron Colembed VL 8B v2 |
|---|---|---|
| Context | 33K | 262K |
| Weights | Not verified | Open |
| Released | — | Jan 2026 |
| Dimensions | — | 4,096 |
| Max input | 33K tokens | 262K tokens |
| Memory (4-bit) | — | 5.2 GB |
| Licence | — | https://huggingface.co/nvidia/nemotron-colembed-vl-8b-v2/blob/main/LICENSE |
| Commercial use | — | Not allowed |
What breaks when you switch
Nemotron 3.5 Embed 8B → Nemotron Colembed VL 8B v2
Re-embed your whole corpus›
Vectors from different models can't be mixed.
Nemotron Colembed VL 8B v2 → Nemotron 3.5 Embed 8B
Re-embed your whole corpus›
Vectors from different models can't be mixed.