What hardware do you need to run Llama 3.3 70B Instruct?

Llama 3.3 70B Instruct at Q4_K_M is 43 GB of weights, with a context ceiling of 128k tokens. A solid generalist. Slower than the 30B MoE models for similar everyday quality. On the cheapest machine that runs it, a Framework Desktop, 128GB, it generates 4 tok/s at 32k of context, measured.

How good is Llama 3.3 70B Instruct, really?

On the Artificial Analysis Intelligence Index v4.3 it scores 8, 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 — good
• Reasoning & maths — usable
• Agentic work — usable

Head to head with its neighbours on the leaderboard: vs DeepSeek-R1-Distill-Llama-70B above it, vs LFM2.5 2.6B below.

Is Llama 3.3 70B Instruct still worth running?

Not on the two figures this site can compare. The intelligence index puts 20 of the current models here above Llama 3.3 70B Instruct, and not one of them asks more of a machine: at 32k of context it needs 53.3 GB and scores 8, where the strongest of the 20, Qwen3.8 27B, needs 18.6 GB and scores 34 (xhigh reasoning effort; 22 without reasoning).

The cheapest machine still sold that holds Llama 3.3 70B Instruct 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 3.3 70B Instruct at 4 tok/s measured. Qwen3.8 27B itself runs on the Framework Desktop, 32GB at $1,269, $2,180 less.

Llama 3.3 70B Instruct's score is Artificial Analysis's own estimate rather than a measured run of the full v4.3 suite.

What Llama 3.3 70B Instruct costs either way

Renting the same model costs $0.1 per million input tokens and $0.32 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 3.3 70B Instruct

Every machine here holds it to 128k, this model's own ceiling, 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 128k 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 27 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 3.3 70B Instruct at 128k on the Strix Halo Framework Desktop, 128GB

How much memory does Llama 3.3 70B Instruct need at 128k?

Every figure above is priced at 32k of context. At 128k, this model's own ceiling, Llama 3.3 70B Instruct needs 85.5 GB rather than 53.3 GB. The weights are the same either way; the key-value cache goes from 10.7 GB to 42.9 GB.

13 of the 37 machines here still sold at a published price hold Llama 3.3 70B Instruct at 32k and 12 hold it at 128k. The cheapest still sold is the same at both lengths: the Framework Desktop, 128GB at $3,449.

On the Framework Desktop, 128GB it runs 4 tok/s measured at 32k and 2.5 tok/s measured at 128k. Every token generated reads the whole cache back, so a longer window is slower as well as bigger.

What has been measured running Llama 3.3 70B Instruct

Two of the speeds in the table above were measured, and there are more measurements than that table shows. Six machines here have a published speed for Llama 3.3 70B Instruct. The table above has two of them; of the other four, two are no longer sold and two are still sold but not the cheapest of their 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. Where two memory tiers of one machine carry the same figure, the speed is set by the bandwidth of the chip they share rather than by the amount of memory. Each published figure links the source that took it.

Four more machines, at a shorter window

The table above lists the machines that run it at 32k, the context the calculator starts on, where Llama 3.3 70B Instruct needs 53 GB. Four more machines hold it at a shorter window, because the weights are the same size either way and the key-value cache is not. The cheapest of those is the Corsair AI Workstation 300, 64GB at $1,700. The table above starts at $3,449.

One machine per family, cheapest first, from the same list as the table above. Each window is the longest setting the calculator offers that the machine still holds this model at, and it opens the calculator on that machine at that length. What it needs there is the weights plus the key-value cache at that window, measured against the memory each machine's GPU can address, which that machine's own page gives.

The specifics

Parameters

70.6B

Quantisation

Q4_K_M

Weights on disk

43 GB

KV cache

11 GB at 32k context. How weights and cache add up.

Maximum context

128k tokens

Licence

Llama 3.3 Community License

Sources

bartowski/Llama-3.3-70B-Instruct-GGUF, unsloth/Llama-3.3-70B-Instruct