Also interesting to see what effort it will take to fine-tune this beast. The latest AISI benchmarks on cybersec place it above glm5.2, but still way way behind SotA closed models. Some fine-tuning might be needed here. Also, interesting to see if Cursor does another training round on it, to directly compare it w/ kimi2.6/2.7 fine-tunes (composer series) and grok4.5.
Also also, interesting to see if someone takes on distilling (proper distillation, w/ training the entire distribution) from this into smaller models. (dsv4-kimi should be really good, since dsv4 is very cheap to serve)
No, you don't. Without training cost you can infer only the marginal cost of serving this kind of models.
Moreover, you don't know the actual size of closed models (what if Fable is a 10T model? What if it's 1T?)
I've made some proof of concept in https://github.com/woct0rdho/transformers5-qwen3.5-recipe . We can finetune Qwen3.5-35B-A3B in 16 GiB VRAM, and DeepSeek-V4-Flash (284B-A13B) in 90 GiB VRAM, without CPU offload. This works well on unified memory machines like Strix Halo.
Even so, larger models like Kimi-K3 still require multiple GPUs and nodes, and there are a lot more to do compare to single-GPU training.
> SemiAnalysis estimates that Anthropic's current blended gross margin has risen to the mid-60% range, with the API business gross margin exceeding 80%
Of course, people will insist "they are lying", "why should we believe them, it's well known they subsidize API pricing", ...
https://newsletter.semianalysis.com/p/anthropic-3q26-profit-...
https://finance.biggo.com/news/02d45650-b569-4d12-b44d-8d6d8...
It's either having a model struggling along with like 5-10 tokens per second on unified memory, or data center cards with hundreds of GB of VRAM consuming like kW of power. It doesn't seem like there's prosumer GPUs with like 180W-250W TDP and 128 GB or 256 GB of VRAM (one can dream).
Obviously nobody is running Kimi K3 locally without an insanely beefy homelab and lots of money to burn, but running GLM 5.2 would be cool at like ~100 tokens per second for a single session and maybe ~60 tokens per second with N subagents.
How unfortunate.
I saw arguments like "Providers cannot price less than their costs" in other comments. In economics, it's generally admitted that they shouldn't price less than their marginal costs, i.e. in their case roughly the cost of electricity, since a lot of these datacenters are not at capacity in terms of graphics cards usage (speculation since it's very easy to rent a GC for a couple hours on some providers). My guess is that someone will be selling tokens at less than electricity + depreciation of GCs soon, since there's a lot of competition and "smaller" data centers have overcapacity? This is speculation, correct me if I'm wrong
The separator is often the only way to distinguish American notation from ISO, so please use a dash for dd-mm-yy and a forward slash for mm/dd/yy
There are quite some countries around the world using d/m/y
https://en.wikipedia.org/wiki/List_of_date_formats_by_countr...
Algeria, Belgium, Brazil, Chile...
We won't be able to run this ourselves, but many providers can.
Hmm, not quite true, I think that honor, for better or worse, goes to OpenAI. When they released GPT2 (or GPT1 for that matter) is was quite literally the SOTA in the ecosystem when it was released.
On the other note, can't wait to see 1bit quantisation soon and how it performs in benchmarks, if it performs really well in benchmarks, would be very good news for GPU hosting providers, to offer "Opus 4.5 level model at the cost of Haiku 4.5"
I wouldn't be surprised if they started offering potentiall bad quantizations with much reduced capability at lower prices (without telling the users, of course)