What hardware do you need to run Llama 4 Scout 17B-16E?

Llama 4 Scout 17B-16E at Q4_K_M is 65 GB of weights, with a context ceiling of 10M tokens. Enormous context on paper. Reception on quality-for-size was mixed; the 30B MoEs are usually the better buy. On the cheapest machine that runs it, a Framework Desktop, 128GB, it generates 16 tok/s at 32k of context, measured.

How good is Llama 4 Scout 17B-16E, really?

On the Artificial Analysis Intelligence Index v4.3 it scores 6, which puts it in the Below every hosted tier band. Fine for simple, well-specified tasks. Noticeably less capable than anything the big labs sell today. See the whole table.

• Summarising — good
• Translation — good
• Everyday coding — usable
• Reasoning & maths — usable
• Agentic work — usable

Head to head with its neighbours on the leaderboard: vs Qwen3 14B above it, vs Ministral 3 14B below. One more model needs much the same memory at 32k of context: vs gpt-oss-120b.

Is Llama 4 Scout 17B-16E still worth running?

Not on the two figures this site can compare. The intelligence index puts 23 of the current models here above Llama 4 Scout 17B-16E, and not one of them asks more of a machine: at 32k of context it needs 68.2 GB and scores 6, where the strongest of the 23, Qwen3.8 27B, needs 18.6 GB and scores 34 (xhigh reasoning effort; 22 without reasoning).

The cheapest machine still sold that holds Llama 4 Scout 17B-16E at 32k is the Framework Desktop, 128GB at $3,449, which holds Qwen3.8 27B as well. This site quotes Qwen3.8 27B at 10 tok/s estimated there and Llama 4 Scout 17B-16E at 16 tok/s measured. Speed is the one figure that goes the other way: Llama 4 Scout 17B-16E is the quicker of the two there. Qwen3.8 27B itself runs on the Framework Desktop, 32GB at $1,269, $2,180 less.

What Llama 4 Scout 17B-16E costs either way

Renting the same model costs $0.1 per million input tokens and $0.3 per million output (OpenRouter, cheapest active endpoint, checked 2026-10-29). Buying a machine only beats that if you use it hard enough, for long enough, that the hardware price divides down below the rental bill. What a million tokens costs each way puts the two prices side by side.

Machines that run Llama 4 Scout 17B-16E

Every machine here holds it to 256k, the longest the calculator offers, so the choice between them is speed and price rather than how much you can keep in the window.

The table is the cheapest machine in each family, not the only one that runs it: 19 of the 56 machines on this site hold it at 32k. Within a family only memory decides, so that is every Mac Studio and NVIDIA card with 96GB or more, every Strix Halo box and MacBook Pro with 128GB or more and the DGX Spark, 128GB. One table sets every machine against the others: price, the memory its GPU can use, what it holds and how long it takes to pay for itself.

One machine per family, cheapest first. Speeds are measured where a public benchmark exists and estimated from memory bandwidth otherwise; the calculator says which for any configuration. The longest context is the longest setting the calculator offers that the machine still holds this model at, cache included, and no machine is shown taking it past its own 10M limit. Each one opens the calculator on that machine at that length.

For scale, the calculator starts from 80 tok/s for a hosted API and times a local machine against it. Nothing above reaches it, and the quickest is 43 tok/s.

The cheapest machine that runs it is sold with 128 GB. There is a page on what 128 GB runs, machine by machine: it names the 11 machines here sold at that size, and the 39 current models the roomiest of them holds at 32k.

It fits in less memory than that: the smallest size here that runs it at 32k is 96 GB, on the RTX PRO 6000 Blackwell, 96GB at $18,000, card only, which hands a model 95 GB.

Run Llama 4 Scout 17B-16E at 256k on the Strix Halo Framework Desktop, 128GB

How much memory does Llama 4 Scout 17B-16E need at 256k?

Every figure above is priced at 32k of context. At 256k, the longest the calculator offers, Llama 4 Scout 17B-16E needs 79.5 GB rather than 68.2 GB. The weights are the same either way; the key-value cache goes from 2.8 GB to 14.1 GB.

13 of the 37 machines here still sold at a published price hold Llama 4 Scout 17B-16E at 32k and 12 hold it at 256k. The cheapest still sold is the same at both lengths: the Framework Desktop, 128GB at $3,449.

On the Framework Desktop, 128GB it runs 16 tok/s measured at 32k and 8.5 tok/s measured at 256k. Every token generated reads the whole cache back, so a longer window is slower as well as bigger.

What has been measured running Llama 4 Scout 17B-16E

One of the speeds in the table above was measured, and there are more measurements than that table shows. Three machines here have a published speed for Llama 4 Scout 17B-16E. The table above has one of them; of the other two, one is no longer sold and one is still sold but not the cheapest of its family. They are below, as published and again at 32k of context, which is what every other speed on this page is taken at.

A published figure is what its source reports, at the window the model was run at. The 32k figure is that measurement read at 32k of context: every token generated reads the weights and the whole key-value cache, so the longer the window the slower it runs. A discontinued machine's price is the one it launched at. Each published figure links the source that took it.

The specifics

Parameters

108.6B, of which 17B are active per token

Quantisation

Q4_K_M

Weights on disk

65 GB

KV cache

2.8 GB at 32k context — 36 of 48 layers use chunked attention with an 8192-token chunk; modelled here as a sliding window. How weights and cache add up.

Maximum context

10M tokens (claimed; nobody has the memory for that locally)

Licence

Llama 4 Community License

Sources

unsloth/Llama-4-Scout-17B-16E-Instruct-GGUF, unsloth/Llama-4-Scout-17B-16E-Instruct