FITMYLLM · NOVEMBER 4, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
74.89B
Parameters (9B active)
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K 64K 128K 256K
Quant Bits VRAM @ 16K Quality IQ2_M 2.93 28.0 GB
27.9 + 0.1 KV
low Q2_K 3.16 30.2 GB
30.1 + 0.1 KV
low IQ3_XXS 3.25 31.0 GB
30.9 + 0.1 KV
low IQ3_XS 3.5 33.3 GB
33.3 + 0.1 KV
low Q3_K_S 3.64 34.7 GB
34.6 + 0.1 KV
low IQ3_M 3.76 35.8 GB
35.7 + 0.1 KV
low Q3_K_M 4 38.0 GB
37.9 + 0.1 KV
low Q3_K_L 4.3 40.8 GB
40.7 + 0.1 KV
moderate IQ4_XS 4.46 42.3 GB
42.2 + 0.1 KV
moderate Q4_K_S 4.67 44.3 GB
44.2 + 0.1 KV
moderate Q4_K_M 4.89 46.4 GB
46.3 + 0.1 KV
good Q5_K_S 5.57 52.7 GB
52.6 + 0.1 KV
good Q5_K_M 5.7 53.9 GB
53.8 + 0.1 KV
good Q6_K 6.56 62.0 GB
61.9 + 0.1 KV
excellent Q8_0 8.5 80.2 GB
80.1 + 0.1 KV
lossless FP16 16 150.4 GB
150.3 + 0.1 KV
lossless
Select your GPU above to see speed estimates and compatibility for each quantization.
Deploying for a team or in production? Size GPUs, cost & scaling in Enterprise → ▸ READY TO RUN THIS? RENT BY THE HOUR
RENT A GPU AND RUN NEMOTRON 3.5 BOLT 74B-A9B NOW
Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.
Community Ratings Chat Coding Reasoning Creative Vision Roleplay Agentic
Loading ratings...
Run this model IQ2_M — 27.9 GB VRAM Q2_K — 30.1 GB VRAM IQ3_XXS — 30.9 GB VRAM IQ3_XS — 33.3 GB VRAM Q3_K_S — 34.6 GB VRAM IQ3_M — 35.7 GB VRAM Q3_K_M — 37.9 GB VRAM Q3_K_L — 40.7 GB VRAM IQ4_XS — 42.2 GB VRAM Q4_K_S — 44.2 GB VRAM Q4_K_M — 46.3 GB VRAM Q5_K_S — 52.6 GB VRAM Q5_K_M — 53.8 GB VRAM Q6_K — 61.9 GB VRAM Q8_0 — 80.1 GB VRAM FP16 — 150.3 GB VRAM
Ollama llama.cpp vLLM LM Studio KoboldCpp Jan Docker
▸ Easiest way to get started · Beginners
DOCS ↗ curl -fsSL https://ollama.com/install.sh | shCOPY
$ ollama run nemotron-3.5-bolt:74b-a9bCOPY
Tag may need adjustment — check ollama.com/library/nemotron-3.5-bolt for available tags.
▸ SETUP GUIDE >_
Auto-setup with fitmyllm CLI Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.
Auto-detect GPU Live tok/s in chat Speed benchmarks 9 inference engines
GPUs that can run this model At Q4_K_M quantization. Sorted by minimum VRAM.
Find the best GPU for Nemotron 3.5 Bolt 74B-A9B
Build Hardware for Nemotron 3.5 Bolt 74B-A9B ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
Nemotron 3.5 Bolt 74B-A9B — 74.89B MoE. ▸ SPECIFICATIONS
PARAMETERS 74.89B (9B active)
ARCHITECTURE Mixture of Experts
CONTEXT LENGTH 256K tokens
CAPABILITIES chat, coding, reasoning, multilingual, vision, agentic, tool_use
RELEASE DATE 2026-10-15
PROVIDER NVIDIA
FAMILY nemotron ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ2_M 2.93 27.9 GB 75% Q2_K 3.16 30.1 GB 78% IQ3_XXS 3.25 30.9 GB 82% IQ3_XS 3.5 33.3 GB 84% Q3_K_S 3.64 34.6 GB 85% IQ3_M 3.76 35.7 GB 86% Q3_K_M 4 37.9 GB 88% Q3_K_L 4.3 40.7 GB 90% IQ4_XS 4.46 42.2 GB 92% Q4_K_S 4.67 44.2 GB 93% Q4_K_M 4.89 46.3 GB 94% Q5_K_S 5.57 52.6 GB 96% Q5_K_M 5.7 53.8 GB 96% Q6_K 6.56 61.9 GB 97% Q8_0 8.5 80.1 GB 100% FP16 16 150.3 GB 100%
§ 01 BENCHMARK SCORES
GPQA Diamond 88.8
HLE 24.7
aa_scicode 39.7
aa_lcr 69.7
aa_terminal_bench_v4 0.0
§ 02 RUN COMMAND
Run Nemotron 3.5 Bolt 74B-A9B locally with Ollama — needs 46.3 GB VRAM at Q4_K_M:
$ ollama run nemotron-3.5-bolt:74b-a9b
§ 03 COMPATIBLE GPUs
30 @ Q4_K_M Feedback 