With RAM you would need at least 500gb to load it but some 100-200gb more for context too. Pair it with a 24gb GPU and the speed will be 10t/s, at least, I estimate.

Oh yes for the FP8, you will need 500GB ish. 4bit around 250GB - offloading MoE experts / layers to RAM will definitely help - as you mentioned a 24GB card should be enough!

Do we know if the full model is FP8 or FP16/BF16? The hugging face page says BF16: https://huggingface.co/Qwen/Qwen3-Coder-480B-A35B-Instruct

So likely it needs 2x the memory.

I think it's BF16 trained then quantized to FP8, but unsure fully - I was also trying to find out if they used FP8 for training natively!

Qwen uses 16bit, Kimi and Deepseek uses FP8.

Oh ok cool thanks!

You know you can discharge a bullet by hitting with a rock right? You don't need a gun.

No, you can discharge a cartridge by hitting it with an improvised device. A bullet is just the metal bit in the front of the cartridge that goes flying. Hitting a bullet with a rock does nothing except maybe dent it.

Even for free users, that is nice

At these prices, I would just get 2xDigits for $6k and have 256gb.

I have a feeling that Digits will probably get sold out and will pricing will get hiked WAY up.

is it confirmed that you can get 256gb of vram for that amount? Because my understanding is that digits pricing will start at $3k for some basic config.

What they meant is buying two whole separate computers.

I understand. It is still unclear if you can get 128GB vram for $3k.

Well, I mean, the press release is pretty unambiguous.

>Each Project DIGITS features 128GB of unified, coherent memory and up to 4TB of NVMe storage.

Even if $3k is only the starting price, it doesn't sound like spending more buys you more memory.

Ok, but it is not clear what kind of RAM is that, how many memory channels, etc. If the goal is to have just 128GB of some ram, then it could be achieved by paying few $100.

Fine, but at that point you're arguing about the concept of the product. It's billed as a computer for AI and you're saying that it might not be more suitable for AI than a regular PC.

it is possible that one could build better PC than digits for AI. We will see once they release digits.

Yep, I think the same. With 128GB fast memory one could run this.

Nice initiative, but I feel the ones who can afford to run R1 locally also have the resource to download it

Yeah, 404 of GPU ram is quite the gpu cluster.

Because RSS is "too old"...

Geo-Guesser future champ right here

Wow what a great read, thank you so much for this! I think every dev should read this at least once

Congratulations Gukesh! Amazing run, truly living a dream.