Quick start
ollama run nemotron3:33bRunnable variants
Memory is a rough file-size estimate; context length, runtime and GPU offload change it.
4 of 4 matching variants shown.4 of 4 variants shown.
| Tag | Type | Inputs | Digest | Download | Parameters | Quantization | Context | Est. memory | Run |
|---|---|---|---|---|---|---|---|---|---|
| 33bdefault | Typelocal | Inputstext, image | Digestf6d8b7ff496c | Download28 GB | Parameters33b | QuantizationUnknown | Context128K | Est. memory≈ 32.5 GB | ollama run nemotron3:33b |
| 33b-bf16 | Typelocal | Inputstext, image | Digestb318f0929784 | Download66 GB | Parameters33b | QuantizationBF16 | Context128K | Est. memory≈ 75 GB | ollama run nemotron3:33b-bf16 |
| 33b-q4_K_M | Typelocal | Inputstext, image | Digestbaa676a14e13 | Download28 GB | Parameters33b | QuantizationQ4_K_M | Context128K | Est. memory≈ 32.5 GB | ollama run nemotron3:33b-q4_K_M |
| 33b-q8 | Typelocal | Inputstext, image | Digest74d89c84a443 | Download36 GB | Parameters33b | QuantizationQ8 | Context128K | Est. memory≈ 41.5 GB | ollama run nemotron3:33b-q8 |
Source-listed applications
Claude CodeHermes AgentOpenClawOpenCode
Related models
nemotron-3.5-boltlocal
Nemotron 3.5 Bolt is NVIDIA’s compact vision-language model for local coding, reasoning, and tool use.
qwen3.8local
Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks.
qwen3.6local
Qwen3.6 delivers substantial upgrades in agentic coding and thinking preservation than previous Qwen models.
muse-glimmerlocal
Meta's latest open model built for always-on local agents. 30B parameters, licensed under Apache 2.0 and runs on a single GPU — tuned for tool use, long tasks, and failure recovery.