Home/Models/Nemotron 3.5 Bolt 74B-A9B
NVIDIA logo

Nemotron 3.5 Bolt 74B-A9B

by NVIDIA · Released Oct 2026

Multimodal
Compare
Context
262K tokens (~131 books)
Input $/1M
n/a
Output $/1M
n/a
Type
multimodal
License
Proprietary
Benchmarks
0 tested
Data as of
About

NVIDIA Nemotron 3.5 Bolt 74B-A9B 24GB is an open vision-language model compressed from Nemotron 3.5 Super VL, with 74B total and 9B active parameters and native 2-bit expert weights so it fits in 24GB for local agents on RTX and DGX Spark. Up to 256K tokens of context. OpenMDW-1.1.

No benchmark data available yet.

Links
Documentation
Community
BenchGecko API
nemotron-3-5-bolt-74b-a9b
Specifications
  • Typemultimodal
  • Context262K tokens (~131 books)
  • ReleasedOct 2026
  • LicenseProprietary
  • StatusActive
Available On
NVIDIA logoNVIDIAn/a
Share & Export
Tweet
Nemotron 3.5 Bolt 74B-A9B is a proprietary multimodal AI model by NVIDIA, released in October 2026. Context window: 262K tokens.

Identifiers and sizes

  • Hugging Face repo · nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB
  • License · openmdw-1.1
  • Total parameters · 24.6B (24,558,610,440) · source Hugging Face safetensors · fetched 2026-11-14
  • Weights size, computed · 49.1 GB at BF16/FP16 · 24.6 GB at INT8/Q8 · 12.3 GB at Q4 (computed from the parameter count, not measured; excludes KV cache and runtime overhead)

Key facts · as of 2026-10-15

  • Nemotron 3.5 Bolt 74B-A9B by NVIDIA. Not enough public benchmark scores to rank yet.
  • List price n/a input · n/a output per 1M tokens (as of 2026-10-15).

How to cite · data as of 2026-10-15

Nemotron 3.5 Bolt 74B-A9B · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-15. https://benchgecko.ai/model/nemotron-3-5-bolt-74b-a9b

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP