MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second • Hacker News Board

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#151

Post by adrian_b » Mon, Jun 08, 2026, 5:16 PM UTC

scosman wrote: post #19

Cerebras is trialing Kimi K2.6 at 3000t/s (invite only). I'm excited for when the fast hardware gets more mainstream for frontier models. Models designed for speed on Nvidia are nice addition that could bridge the gap.

TFA mentions that until now special very expensive hardware like Cerebras was required for reaching this kind of speeds, and it emphasizes that what is novel in their results is that they have obtained over 1000 token/s for a model with over 1 T parameters by using just standard hardware, i.e. one server with 8 GPUs.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#152

Post by ttoinou » Mon, Jun 08, 2026, 5:16 PM UTC

dakiol wrote: post #136

So, regarding the productivity argument: I don't get it. It doesn't really matter (for regular employees) that you can do now in 2h what before it took 2 days. Why? Because it's not that you have the rest of the day for yourself. You still have to work 8h/day as usual. But now the pattern is different: instead of enjoying the craft digging deeper into problems in the span of 2 days, now you are rushing into some slot…

In which world do you live where employees work 8 hours per day ? They clock 8 hours per day maybe, but they don't work that time

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#153

Post by drawfloat » Mon, Jun 08, 2026, 5:16 PM UTC

ignoramous wrote: post #72

miroljub wrote: post #61

Earlier quoted context omitted.

MiMo and DeepSeek are not cheap. Anthropic and OpenAI are expensive for what they provide.

The Chinese "Neijuan" is real & well reported: https://www.reuters.com/business/autos-transportation/what-i... It is another thing the BigLabs accuse open weight models of benefiting from distillation & other techniques & essentially avoid higher training costs (which typically bleed into bills end users pay for inference). Ex A: https://www.anthropic.com/research/2028-ai-leadership Ex B: https://www.reuters.com/worl…

We just had years of US model providers arguing it was fine to rip off the world’s cultural output for their own profit, why should their work be treated any different?

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#154

Post by 0xbadcafebee » Mon, Jun 08, 2026, 5:17 PM UTC

This is the value prop of Groq and Cerebras. They don't have the best models, but they have the fastest inference, and Groq has both the lowest cost and fastest speed.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#155

Post by throwaway67678 » Mon, Jun 08, 2026, 5:18 PM UTC

ecshafer wrote: post #132

throwaway67678 wrote: post #73

Earlier quoted context omitted.

Lower cost of labor, lots of under the hood optimizations (e.g. cache hits for DS), many of these companies have existing infra (fewer upfront costs for deployment), etc

China isn't that cheap for labor. And if you think the guys in Z.ai or xiaoxiao aren't the exact same guys from Tsinghua, Peking, MIT, Stanford, CMU, etc. and pulling in amazing salaries you'd be wrong.

I'd assume there's more to the cost of labor than the salaries of the elite folks who do the R&D, but fair point

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#156

Post by btian » Mon, Jun 08, 2026, 5:18 PM UTC

scosman wrote: post #19

Cerebras is trialing Kimi K2.6 at 3000t/s (invite only). I'm excited for when the fast hardware gets more mainstream for frontier models. Models designed for speed on Nvidia are nice addition that could bridge the gap.

Source? Their website says 1000t/s https://www.cerebras.ai/blog/which-is-faster-gemini-3-5-flas...

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#157

Post by andriy_koval » Mon, Jun 08, 2026, 5:18 PM UTC

epolanski wrote: post #94

alkyon wrote: post #44

Earlier quoted context omitted.

Sounds like exponential growth of crappy software. I'm not saying that before we didn't have mass produced crap in SE, but now it will turn into explosive overflow.

I am more and more inclined into not believing this crappy software theory. Especially as teams invest in proper agentic harnessing. We have had a champion in our team that has invested a lot of time into it over the last 4 months, and if anything, quality has improved, not decreased. Architecture is more coherent, codebase has been cleaned up, agents find information quickly, code produced is very solid and my role…

> We have had a champion in our team

there are good actors, which are empowered by AI to produce positive impact, but often there are N times more bad actors, which push crappy code to close feature requests fast, increase performance LoC-like metrics, etc.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#158

Post by pants2 » Mon, Jun 08, 2026, 5:19 PM UTC

With a tps and a token price you can calculate approx. price per hour of running the model!

$2.61/M tokens * 1,000 tok/s = $9.40/hr

That would be pretty cheap for an 8-GPU node which would typically run around $45/hr or more. Guess this depends on how many parallel streams it can handle.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#159

Post by noncoml » Mon, Jun 08, 2026, 5:19 PM UTC

dakiol wrote: post #136

So, regarding the productivity argument: I don't get it. It doesn't really matter (for regular employees) that you can do now in 2h what before it took 2 days. Why? Because it's not that you have the rest of the day for yourself. You still have to work 8h/day as usual. But now the pattern is different: instead of enjoying the craft digging deeper into problems in the span of 2 days, now you are rushing into some slot…

You have to think LLM as the genie that tries to trick you.

First make it write a contract (REQ/ARCH/IMPL documents). Skim through those for any mistakes.

Then based on those ask it to write tests. Again skim through them.

Now you have a context full of guardrails. It’s less likely to surprise you.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#160

Post by altcognito » Mon, Jun 08, 2026, 5:21 PM UTC

epolanski wrote: post #94

alkyon wrote: post #44

Earlier quoted context omitted.

Sounds like exponential growth of crappy software. I'm not saying that before we didn't have mass produced crap in SE, but now it will turn into explosive overflow.

I am more and more inclined into not believing this crappy software theory. Especially as teams invest in proper agentic harnessing. We have had a champion in our team that has invested a lot of time into it over the last 4 months, and if anything, quality has improved, not decreased. Architecture is more coherent, codebase has been cleaned up, agents find information quickly, code produced is very solid and my role…

It makes no sense. I mean, T2 covered this:

"Watching John with the machine, it was suddenly so clear. The terminator would never stop. It would never leave him, and it would never hurt him, never shout at him, or get drunk and hit him, or say it was too busy to spend time with him. It would always be there. And it would die to protect him. Of all the would-be fathers who came and went over the years, this thing, this machine, was the only one who measured up. In an insane world, it was the sanest choice."

As long as you've indicated what you want, the machine will try to do what you ask of it. It won't get tired because "the codebase is too big", or it has gotten bored of the pattern, or it wants to introduce a new technology.

It just does the thing you asked of it. (note, that yes, I get that as a codebase size increases, it might make it more difficult to fit into context, but that only applies if it needs to read a large percentage of the project to implement the task, which shouldn't be the case.