Top
Best
New

Posted by Retro_Dev 15 hours ago

Mercury 2.5 LLM hits 770 tokens per second(artificialanalysis.ai)
133 points | 77 comments
jjcm 9 hours ago|
I'm still sad that we haven't seen a new Taalas style chip a la https://chatjimmy.ai/. Smaller models are good enough now to make that insane burst of tokens so useful.
pil0u 4 hours ago||
I don't know the model behind this, but it is absurdly bad.

> Write me a coherent paragraph in French, without ever using the letter "e".

> Voilà une phrase claire et concise : "Le village est situé dans les montagnes. Le soleil est haut. Il y a des animaux dans le village. Il pleut dans les montagnes."

I suppose this is just a demo of how fast an LLM can be, I wonder if there are tradeoffs with larger/smarter models. Also, for a human usage, at what point are tokens generated fast enough that it's pretty much instant? My bet is below 1000 tps

PetahNZ 4 hours ago|||
Its Llama 3.1 8B, a very old/small model.
kgeist 1 hour ago|||
That's the problem with etching a model onto a chip: by the time you've designed the chip, manufactured it, tested it, shipped it, and deployed it, the model will be hopelessly outdated (with the current improvement rates). And when you want to update, you have to buy new chips instead of just uploading a new model file like now. When Taalas announced their chip, the model was already 1.5 years old (stone age by current standards). It's their first chip, so maybe they can streamline it, but the problem of having to update hardware every few months to keep up with the industry is not going anywhere.
dnautics 19 minutes ago|||
You can upload different weights and even do LoRAs. The chip architecture is interesting, the first (n) layers are the sane, so you can change architecture by adding (m) layers. Plausible that this is sufficiently flexible enough for several generations of real world applications. For example, we still use 45nm general purpose silicon for automotive, e.g.
linuxftw 19 minutes ago||||
What they did had never been done before. Now we see that it's possible, there are plenty of models to choose from that could be etched into silicon. In the next year or two, I think these smaller models might plateau, and there may be some on-device niche they can fill.
busssard 35 minutes ago|||
[dead]
carlos_rpn 1 hour ago||||
Wasn't it also quantized aggressively, like 1 or 2 bits?
cyanydeez 3 hours ago|||
for comparison, Qwen3.8-Flash-Next only requires 6B parameters for computation, but stores 125B, 51B of those can be comfortably offloaded as they're not actively used in decode but a single token look up.

The Quant iQ4 of this model loads, then, in ~60GB of vram, and on disk it's 85GB.

So if you could etch it, you'd need a ~25GB ssd chip and 60GB of vram.

The vram costs likely contributed to these things being out of reach of the current economic cycle.

Tuna-Fish 2 hours ago||
You don't want to use a sparse model for a Taalas-like design. Something like a Qwen 3.8 27B makes much more sense.
fransje26 4 hours ago|||
Then again, good luck writing a coherent paragraph in French without an "e". :-)
mejutoco 2 hours ago|||
There is a book written under this premise. Probably the inspiration for that prompt

https://en.wikipedia.org/wiki/A_Void

winwang 8 hours ago||
Pretty sure context is in SRAM, and then you have that as a blocker for tasks.
electroglyph 5 hours ago||
way less of a blocker these days due to sparse attention...
bearjaws 14 hours ago||
If you care about speed Cerebras gpt-oss-120b is 1400tk/s and "just as smart" in ranking.

I've used it on a few for fun projects and its decent but the speed is crazy to watch.

ford 13 hours ago||
Also kimi 2.6 at 1000tps (as of may), though when we reached out they had a >12 month waitlist and minimum 7-8 figure annual token spend.

[0] https://www.cerebras.ai/blog/cerebras-kimi-k2-Enterprise

walrus01 10 hours ago|||
7-8 figures annual spend will buy a hell of a lot of capable local inference hardware you can own, though it won't be at the absurd token/s rate, you'll be able to run almost anything on it... And it'll still have a good residual resale value after 4 years the way things are going now.
podocarp 6 hours ago||
Feel like you could spend 6 figures building out a team and the rest renting compute for a whole year, and get the team to create a local inference solution with that kind of budget…
sharktheone 12 hours ago|||
yeah. K2.6 can run on insane speeds. So sad that they don't have K3 yet.

But it can apparently also run 5.6 Sol

eli 11 hours ago||
Yeah but at “call to discuss pricing” rates
conception 10 hours ago|||
It was better when they had gemma at 1k. Inco does DS flash at about 600. A few places will do K3 and GLM in the hundreds.

Such a tiny model at that t/s is less impressive than it would have been four months ago.

physicallyIllfr 9 hours ago|||
Lighting my codebase on fire at the speed of light. Like microwaving the spaghetti.

I genuinly only see these speeds being useful for customer service/transactional workflows. Of which much smaller models can do the job (but those dont make tons of money for companies like Cerebras that need to pay off massive amounts of debt).

Nobody needs to code at 600 words per second. Using a 100tps model for an hour or so will leave you with 4-8hrs of code review and revision work.

sbierwagen 7 hours ago|||
Human code review? What is this, 2025? The modality today is write with one LLM, review by a different one, (important: two different model families will catch errors one series won't) then deploy right to production.
pbasista 20 minutes ago|||
> then deploy right to production

And when things go south, you blame AI?

I would be very curious to see how you explain it to your customers.

Is it going to sound similar to this?

> You see, our well-meaning AI-generated code has caused all your data to be permanently deleted. In case you are confused as to who to blame, we would like to clarify that we did not write, nor review the code. So we cannot possibly bear any responsibility for its mistakes. The responsibility lies with the LLMs, not us. We have already fired the LLM which did the coding and the review. And we are already using their main competitors. Hopefully that settles your concern with the quality of our service and we are looking to have you on board of our next products.

podocarp 6 hours ago|||
And if you tell the reviewer the author is a competitors model it becomes extra snarky and vigilant. Then give the review results to the author and tell it it's from the competition and it will also become slightly outraged.
calgoo 5 hours ago||||
No you dont need to code at 600 w/s BUT at those speeds, you can start doing things like asking multiple different agents the same question and picking the best solution each time without noticing the lag.
RugnirViking 5 hours ago|||
> Nobody needs to code at 600 words per second.

I do. I used to use haiku for the speed. Now its just as slow as the rest. Speed is my #1 ranking of how good a model is

lostmsu 2 hours ago|||
Inco sucks. I tried their GLM 5.3 Flash and it was quantized to the point of hallucinating Chinese in the middle of English only agentic sessions. Never happened with any other provider.
LoganDark 12 hours ago|||
Please do not try to use gpt-oss-120b over Cerebras. It is broken, screws up tool calls most of the time, forgets to end thinking blocks and has all sorts of other issues. The speed is amazing but it is absolutely not worth it, especially at that quite incredible cost. Think: $5–10/minute levels of cost with a single agent, because Cerebras also offers no cache pricing for input tokens at all.
bearjaws 1 hour ago|||
Not been my experience, I have it using tool calls in a video game I am building and it correctly adheres ~99% of the time.

I have it retry on failure, but you should do that with any LLM really.

LoganDark 1 hour ago||
I kept having experiences with gpt-oss-120b on Cerebras where it would get stuck in a thinking block and then start endlessly saying things like "Running the command now." or "Making the changes now." and then simply repeating similar sentences like that forever instead of actually making the tool call. It made tool calls other times, so it wasn't an issue with tool calls being impossible, but it just wasn't doing a good job of using them for real instead of simply saying it would. So this was not an issue of it starting a tool call and then putting invalid syntax inside of it, it just would not make the tool call it was supposed to whatsoever. There's no automatic way to retry that.
eli 11 hours ago||||
Which is wild because it does, in fact, do caching
fakwandi_priv 7 hours ago||
> There is no additional fee for using prompt caching. Input tokens, whether served from the cache or processed fresh, are billed at the standard input token rate for the respective model.

So what he’s saying is correct, there is no separate cache pricing, which by normal standards should be 10% of the cost, which can become exceedingly expensive for anything other than single turn. The way they are stating this is of course strange..

shard972 10 hours ago|||
Yea i had some pretty meh results using gpt-oss-120b it in my evals where it should have benefited speed alot but it really under performed what i was expecting.
scosman 12 hours ago|||
Or better: Qwen 2.8 27b
RussianCow 12 hours ago||
Unfortunately, the lack of an input cache discount makes it prohibitively expensive for most use cases that aren't one-shot prompts.
scosman 9 hours ago||
well same applies to GPT OSS 120. Qwen is just the much smarter model of the 2 public options on Cerebras.
kylehotchkiss 8 hours ago|||
I felt happy I could run it at 100tk/s on my new Mac Studio :')
verdverm 9 hours ago||
if you want to feel the speed without the burn

https://kamilstanuch.github.io/LLM-token-generation-simulato...

the_arun 9 hours ago||
Chat Jimmy clocks at 17K tokens per sec burning LLM into the Chip - https://chatjimmy.ai/ - Source: https://theashishmaurya.medium.com/taalas-the-startup-that-p...
econ 9 hours ago|
Would have made a nice local tool.
freakynit 10 hours ago||
I have tried using Mercury 2.5 for a lot of my tasks.. but this model just isn't there. It seems to be on par with any 14B model at max. Even GPT-OSS-20B performs way better than this in my own attempts to use it.

I really really wanted to use this because it offers incredible speeds and pricing combinations. But nop.. I still am not using it.. not even for basic tasks.

nostrebored 10 hours ago|
Try Celeris-magnus-1. We get similar speeds and it’s much closer to qwen 27B dense models.
freakynit 9 hours ago||
Didn't knew about this. Thanks.. but, according artificialanalysis.ai, it's intelligence is just about like a 14B model (mistral 3 14b)
nostrebored 7 hours ago||
depends on your use case I think. for small, transactional tasks it is quite strong and fast.
walrus01 14 hours ago||
Pricing at $0.25 and $0.75 already puts its cost well above reasonably reputable inference providers for deepseek v4 flash or qwen 3.8-flash-next or similar class of open weight LLMs that fit in under 170GB of RAM, so I don't see the point. I think this is probably also stupider than laguna s 2.1 which can also be very cheap to serve.
RussianCow 12 hours ago||
The point is the speed.
_aavaa_ 11 hours ago||
If the model provides me with bad results because it's dumb, I don't care how quickly it does it.
RussianCow 8 hours ago|||
But there are lots of use cases where a relatively "dumb" model is good enough.
hermannj314 10 hours ago||||
Is it possible to construct a control system where bad, fast and cheap can become good, fast, and cheap through repeated sampling and a strong spec/eval harness?

I am trying to keep an open mind with AI, but I also have little understanding of control theory, trying to learn.

calgoo 5 hours ago|||
You can, but you need to break the problem into much smaller tasks, then check those answers, and finally have a harness that handles all the context, task breakup, task definitions, and validations each round.
freakynit 10 hours ago|||
Smaller models seems to get stuck in "loops" when you try to "handle" them this way.
jgalt212 1 hour ago|||
fast results that you need to verify are better than slow (allegedly better) results that you still need to verify. REPL vs batch.
nextaccountic 10 hours ago||
At some point the bottleneck becomes tool calling.. and as such, it's preferably if the model is co-hosted (in the same datacenter, at least) with your code repository and all other reference/context it needs (full documentation for most ecosystems, maybe even a copy of common crawl to minimize web fetch usage, etc)
nylonstrung 14 hours ago||
I honestly think the diffusion LLM approach is a dead end

It's telling that frontier labs like Google toyed around with it but didn't invest further even for their most speed and cost sensitive small models

Still unclear for what, if any use cases this is pareto frontier

vineyardmike 11 hours ago||
From what I’ve heard, the issue is more that it’s harder to efficiently share the hardware across diffusion requests, so it’s more expensive to serve.

It sounds like there might be opportunities for local models (not open weight, but actually locally run) to use diffusion for faster responses on weaker hardware that doesn’t need to be shared.

But yea, it’s still a red-ish flag that big labs haven’t invested much in it. I could see Google/Apple getting value of this sort of local model, but maybe there’s enough research behind traditional models that it’s not worth the distraction at this point in time.

LarsDu88 13 hours ago|||
You can't think that a small startup versus Anthropic's training setup is anywhere near the same scale to make apples to apples comparisons.

Not sure how the Chinese labs pull it off though using autoregressive models. The secret sauce is probably going to be in the training data.

The main reason Google hasn't switched over to DiffusionGemma is because serving at larger batch sizes loses the speed gains you get from diffusion, and most of the primary use case is serving many users at once off a single device with a large batch size.

If you were to move to on-device low latency... like say in a robot or something, then the story might be different...

clhodapp 12 hours ago|||
Personally, I think it's more that text diffusion is not the ideal driver of an agentic work loop than that text diffusion is a total dead end. I am still hoping to see how it does on authoring and editing with further scaling and optimization. I think the push for AGI has put a bit too much focus on the idea of one general model doing everything.
naasking 31 minutes ago|||
There were too many unsolved problems with diffusion to commercialize it, where autoregressive loops had a more straightforward roadmap. Doesn't mean the diffusion problems are insurmountable though.
kennywinker 11 hours ago||
DiffusionGemma was released alongside the other gemma-4 models just a few months ago, so clearly google hasn't abandoned the idea.

K2-Horizon-7B has a diffusion and non-diffusion variant, and they claim the same level of intelligence from both models.

big-chungus4 6 hours ago||
My 1k parameter RNN is faster
seduerr 10 hours ago||
Cerebras is fast…?
Retro_Dev 9 hours ago|
Yep it's fast alright. I wonder how much faster Mercury would become if hosted on Cerebras.
low_tech_punk 13 hours ago|
it's stupid fast!
SwellJoe 12 hours ago|
It's stupid and it's fast.
freakynit 10 hours ago|||
Couldn't have said it better.
verdverm 8 hours ago||
but can it be said in fewer tokens than GGP commenter?
supercoder88 10 hours ago|||
[dead]
More comments...