FITMYLLM · SEPTEMBER 6, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
398B
Parameters (94B active)
0
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K 64K 128K 256K
Quant Bits VRAM @ 16K Quality IQ2_XXS 2.38 122.3 GB
118.9 + 3.4 KV
low IQ2_M 2.93 149.6 GB
146.3 + 3.4 KV
low Q2_K 3.16 161.1 GB
157.7 + 3.4 KV
low IQ3_XXS 3.25 165.6 GB
162.2 + 3.4 KV
low IQ3_XS 3.5 178.0 GB
174.6 + 3.4 KV
low Q3_K_S 3.64 185.0 GB
181.6 + 3.4 KV
low IQ3_M 3.76 190.9 GB
187.5 + 3.4 KV
low Q3_K_M 4 202.9 GB
199.5 + 3.4 KV
low Q3_K_L 4.3 217.8 GB
214.4 + 3.4 KV
moderate IQ4_XS 4.46 225.7 GB
222.4 + 3.4 KV
moderate Q4_K_S 4.67 236.2 GB
232.8 + 3.4 KV
moderate Q4_K_M 4.89 247.1 GB
243.8 + 3.4 KV
good Q5_K_S 5.57 281.0 GB
277.6 + 3.4 KV
good Q5_K_M 5.7 287.4 GB
284.1 + 3.4 KV
good Q6_K 6.56 330.2 GB
326.8 + 3.4 KV
excellent Q8_0 8.5 426.7 GB
423.4 + 3.4 KV
lossless FP16 16 799.9 GB
796.5 + 3.4 KV
lossless
Select your GPU above to see speed estimates and compatibility for each quantization.
Too big for a single GPU — plan a multi-GPU deployment
Even the lightest quant needs ~122 GB. Size GPUs, replicas, TCO and scaling for a production setup. Open in Enterprise →
▸ READY TO RUN THIS? RENT BY THE HOUR
RENT A GPU AND RUN JAMBA LARGE 1.6 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_XXS — 118.9 GB VRAM IQ2_M — 146.3 GB VRAM Q2_K — 157.7 GB VRAM IQ3_XXS — 162.2 GB VRAM IQ3_XS — 174.6 GB VRAM Q3_K_S — 181.6 GB VRAM IQ3_M — 187.5 GB VRAM Q3_K_M — 199.5 GB VRAM Q3_K_L — 214.4 GB VRAM IQ4_XS — 222.4 GB VRAM Q4_K_S — 232.8 GB VRAM Q4_K_M — 243.8 GB VRAM Q5_K_S — 277.6 GB VRAM Q5_K_M — 284.1 GB VRAM Q6_K — 326.8 GB VRAM Q8_0 — 423.4 GB VRAM FP16 — 796.5 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 jamba:398b-q4_K_MCOPY
Tag may need adjustment — check ollama.com/library/jamba 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 Jamba Large 1.6
Build Hardware for Jamba Large 1.6 ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
Jamba Large 1.6 — 398B MoE. ▸ SPECIFICATIONS
PARAMETERS 398B (94B active)
ARCHITECTURE Mixture of Experts
CONTEXT LENGTH 256K tokens
CAPABILITIES chat, reasoning, tool_use, multilingual
RELEASE DATE 2025-09-20
PROVIDER AI21
FAMILY jamba ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ2_XXS 2.38 118.9 GB 65% IQ2_M 2.93 146.3 GB 75% Q2_K 3.16 157.7 GB 78% IQ3_XXS 3.25 162.2 GB 82% IQ3_XS 3.5 174.6 GB 84% Q3_K_S 3.64 181.6 GB 85% IQ3_M 3.76 187.5 GB 86% Q3_K_M 4 199.5 GB 88% Q3_K_L 4.3 214.4 GB 90% IQ4_XS 4.46 222.4 GB 92% Q4_K_S 4.67 232.8 GB 93% Q4_K_M 4.89 243.8 GB 94% Q5_K_S 5.57 277.6 GB 96% Q5_K_M 5.7 284.1 GB 96% Q6_K 6.56 326.8 GB 97% Q8_0 8.5 423.4 GB 100% FP16 16 796.5 GB 100%
§ 03 COMPATIBLE GPUs
5 @ Q4_K_M Feedback