NVIDIA Nemotron 3.5 Bolt
74B total / 9B active multimodal model with a distinct native 2-bit 24GB checkpoint and up to 256K context.
- Practical guide
Access choices, version boundaries and evaluation limits.
- Official model source
Verify the exact checkpoint or provider contract.
- Official model source
Verify the exact checkpoint or provider contract.
Model Specificationsnvidia
24GB packed-variant target at batch size one / 128K; BF16 and larger contexts need separate capacity.
Downloadable weights; self-hosting costs vary with hardware, context and concurrency.
Scoped Terms🚀Function Overview
Primary-source contract checked October 21, 2026. No independent model request or hardware test was executed.
Key Features
- Text, image and video input; text output; up to 256K combined context.
- The 24GB packed checkpoint is distinct from NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16.
- The card targets below 24GB at batch size one and 128K context; this is not a memory guarantee at 256K.
- vLLM instructions name Ampere A100, Hopper H100 and Blackwell; confirm your actual runtime and hardware.
- October 15 model-card release; OpenMDW-1.1 license.
Use Cases
- •Compare the exact checkpoint and provider contract
- •Evaluate a permitted task with a verifiable reference
⚙️Input Parameters
model
stringnvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB — follow the selected official runtime or provider interface; this is not a universal hosted alias.
task_input
provider-specificUse the documented input schema; see official sources for supported modalities and request fields.
Quick Actions
Technical Specifications
- Hardware Type
- 24GB packed-variant target at batch size one / 128K; BF16 and larger contexts need separate capacity.
- Commercial Use
- OpenMDW-1.1; commercial and non-commercial use subject to governing terms.
- Pricing
- Downloadable weights; self-hosting costs vary with hardware, context and concurrency.