Smaller alternatives
Models that need less memory than Llama 4 Scout 17B 16E and score about as well or better.
Llama 4 Scout 17B 16E
MoEVision
Meta · 108.6B (17B active) · Mixture of experts
Llama 4 Scout 17B 16E is a 108.6B model from Meta with a 10M-token context. At Q4_K_M with an 8K context it needs about 62.7 GB; 80 GB leaves room for longer chats.
Mixture of experts
Parameters: 108.6B · Active: 17B
All 108.6B parameters load into memory, but only 17B work on each token, so it runs at the speed of a much smaller model.
Quantization options
| Quant | Bits | Memory | Quality | On your hardware |
|---|---|---|---|---|
| Q2_K | 3.16 | — | Noticeably worse | … |
| Q3_K_M | 4 | — | Some loss | … |
| Q4_K_M | 4.89 | — | Good | … |
| Q5_K_M | 5.7 | — | Good | … |
| Q6_K | 6.56 | — | Near original | … |
| Q8_0 | 8.5 | — | Near original | … |
| F16 | 16 | — | Original | … |
Memory by context length
At Q4_K_M. The context cache grows with every token the model keeps in mind.
| Context | Context cache | Total |
|---|---|---|
| 4K | 0.8 GB | 61.9 GB |
| 8K | 1.5 GB | 62.7 GB |
| 32K | 2.6 GB | 63.8 GB |
| 128K | 7.1 GB | 68.3 GB |
| 256K | 13.1 GB | 74.3 GB |
Next steps
Compare it with
Scored models of a similar size, side by side on your device.
Least hardware that runs it well
Desktop devices with the least memory that give Llama 4 Scout 17B 16E grade A or S at Q4_K_M.
Can I run Llama 4 Scout 17B 16E locally?
Can I run Llama 4 Scout 17B 16E locally?
Yes, if your GPU or Mac has about 62.7 GB free for it at Q4_K_M. Open this page on that computer to see the grade for your exact hardware, then start it with llama.cpp, Ollama or LM Studio.
How much memory does Llama 4 Scout 17B 16E need?
About 62.7 GB at Q4_K_M with an 8K context and 108.5 GB at Q8_0. Lower quantizations fit smaller cards with some loss in quality; longer contexts add to the total.