What hardware do you need to run Flint 40B-A2.5B?

Flint 40B-A2.5B at UD-Q4_K_XL is 25 GB of weights, with a context ceiling of 256k tokens. Recursive built it for coding agents, to read a codebase, fix bugs and write tests: 39.8B total with about 2.5B of them active for every token, so a token reads a fraction of the 25 GB on disk. Recursive serves it free through OpenRouter today, and nobody sells it by the token. On the cheapest machine that runs it, a Radeon AI PRO R9700, 32GB, it generates 131 tok/s at 32k of context, worked out from memory bandwidth.

How good is Flint 40B-A2.5B, really?

It has not been placed on the intelligence index yet. See the whole table.

What Flint 40B-A2.5B costs either way

Nobody rents Flint 40B-A2.5B by the token. The closest hosted match, Qwen3.6 35B-A3B, costs $0.05 per million input tokens and $0.7 per million output (OpenRouter, cheapest active endpoint, checked 2026-10-23). 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 Flint 40B-A2.5B

Every machine here runs it, but not to the same length. The Mac Studio M5 Max, 36GB stops at 64k; the Strix Halo Corsair AI Workstation 300, 64GB takes it to 256k, which is this model's own ceiling and the longest the calculator offers. The weights are the same size on every machine; what differs is the memory left for the key-value cache, which grows with every token you keep.

The table is the cheapest machine in each family, not the only one that runs it: 38 of the 56 machines on this site hold it at 32k. Within a family only memory decides, so that is every Mac Studio at any size, every NVIDIA card with 32GB or more, every Mac mini and MacBook Pro with 48GB or more, every Strix Halo box with 64GB or more, the Radeon AI PRO R9700, 32GB 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 256k 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. 2 of the 7 above reach it, and the slowest is 36 tok/s.

The cheapest machine that runs it is sold with 32 GB. There is a page on what 32 GB runs, machine by machine: it names the 6 machines here sold at that size, and the 30 current models the roomiest of them holds at 32k, this one among them.

Run Flint 40B-A2.5B at 256k on the Strix Halo Corsair AI Workstation 300, 64GB

How much memory does Flint 40B-A2.5B need at 256k?

Every figure above is priced at 32k of context. At 256k, which is this model's own ceiling and the longest the calculator offers, Flint 40B-A2.5B needs 32.6 GB rather than 26 GB. The weights are the same either way; the key-value cache goes from 1 GB to 7.6 GB.

27 of the 37 machines here still sold at a published price hold Flint 40B-A2.5B at 32k and 24 hold it at 256k. The cheapest still sold goes from the Radeon AI PRO R9700, 32GB at $1,299, card only to the Corsair AI Workstation 300, 64GB at $1,700. The Radeon AI PRO R9700, 32GB is priced as the card alone, so add the PC around it before comparing it with a complete computer.

On the Corsair AI Workstation 300, 64GB it runs 55 tok/s estimated at 32k and 15 tok/s estimated at 256k. Every token generated reads the whole cache back, so a longer window is slower as well as bigger.

The specifics

Parameters

39.8B, of which 2.5B are active per token

Quantisation

UD-Q4_K_XL

Weights on disk

25 GB

KV cache

1.0 GB at 32k context — 35 of the 49 layers cache 513 tokens and stop growing there; the other 14 cache the whole context, so the full 256k window costs about 7.6 GB. The four-bit build priced here is an unsloth dynamic quant, which allocates bits differently from a plain Q4_K_M; Recursive publishes BF16 and its own four-bit GGUF, which is 26.0 GB. How weights and cache add up.

Maximum context

256k tokens

Licence

Apache 2.0

Sources

RecursiveSI/flint, unsloth/flint-GGUF