One in all Europe’s most distinguished AI startups has launched two AI fashions which can be so tiny, they’ve named them after a hen’s mind and a fly’s mind.
Multiverse Computing claims these are the world’s smallest fashions which can be nonetheless excessive performing and might deal with chat, speech, and even reasoning in a single case.
These new tiny fashions are supposed to be embedded into web of issues units, in addition to run regionally on smartphones, tablets, and PCs.
“We will compress the mannequin a lot that they will match on units,” Orús instructed TechCrunch. “You’ll be able to run them on premises, instantly in your iPhone or in your Apple Watch.”
As we beforehand reported, Multiverse Computing is a buzzy European AI startup headquartered in Donostia, Spain, with about 100 staff in places of work worldwide. It was co-founded by a high European professor of quantum computer systems and physics, Román Orús; quantum computing professional Samuel Mugel; and Enrique Lizaso Olmos the previous deputy CEO of Unnim Banc.
It simply raised €189 million (about $215 million) in June on the power of a mannequin compression expertise it calls “CompactifAI.” (Because it was based in 2019, it has raised about $250 million, Orús mentioned.)
CompactifAI is a quantum-inspired compression algorithm that reduces the scale of current AI fashions with out sacrificing these fashions’ efficiency, Orús mentioned.
Techcrunch occasion
San Francisco
|
October 27-29, 2025
“We now have a compression expertise that’s not the standard compression expertise that the individuals from pc science or machine studying will do, as a result of we come from quantum physics,” he described. “It’s a extra refined and extra refined compression algorithm.”
The corporate has already launched an extended listing of compressed variations of open-source fashions, particularly common small fashions like Llama 4 Scout or Mistral Small 3.1. And it simply launched compressed variations of OpenAI’s two new open fashions. It has additionally compressed some very giant fashions – it provides a DeepSeek R1 Slim, as an example.
However because it’s within the enterprise of creating fashions smaller, it has targeted additional consideration on making the smallest but strongest fashions doable.
Its two new fashions are so small that they will carry chat AI capabilities to only about any IoT system and work with out an web connection, the corporate says. It humorously calls this household the Mannequin Zoo as a result of it’s naming the merchandise based mostly on animal mind sizes.
A mannequin it calls SuperFly is a compressed model of Hugging Face’s open-source mannequin SmolLM2 135. The unique has 135M parameters and was developed for on-device makes use of. SuperFly is 94M parameters, which Orús likens to the scale of a fly’s mind. “That is like having a fly, however slightly bit extra intelligent,” he mentioned.
SuperFly is designed to be skilled on very restricted information, like a tool’s operations. Multiverse envisions it embedded into dwelling home equipment, permitting customers to function them with voice instructions like “begin fast wash” for a washer. Or customers can ask troubleshooting questions. With slightly processing energy (like an Arduino), the mannequin can deal with a voice interface, as the corporate confirmed in a dwell demo to TechCrunch.
The opposite mannequin is called ChickBrain, and is bigger at 3.2 billion parameters, however can be way more succesful and has reasoning capabilities. It’s a compressed model of Meta’s Llama 3.1 8B mannequin, Multiverse says. But it’s sufficiently small to run on a MacBook, no web connection required.
Extra importantly, Orús mentioned that ChickBrain truly barely outperforms the unique in a number of customary benchmarks, together with the language-skill benchmark MMLU-Professional, math expertise benchmarks Math 500 and GSM8K, and the final information benchmark GPQA Diamond.
Listed below are the outcomes of Multiverse’s inside assessments of ChickBrain on the benchmarks. The corporate didn’t provide benchmark outcomes for SuperFly however Multiverse additionally isn’t concentrating on SuperFly at use circumstances that require reasoning.
It’s necessary to notice that Multiverse isn’t claiming that its Mannequin Zoo will beat the most important state-of-the-art fashions on such benchmarks. Zoo performances may not even land on the leaderboards. The purpose is that its tech can shrink mannequin dimension with out a efficiency hit, the corporate says.
Orús says the corporate is already in talks with all of the main system and equipment makers. “We’re speaking with Apple. We’re speaking with Samsung, additionally with Sony and with HP, clearly. HP got here as an investor within the final spherical,” he mentioned. The spherical was led by well-known European VC agency Bullhound Capital, with participation from quite a lot of others, together with HP Tech Ventures and Toshiba.
The startup additionally provides compression tech for different types of machine studying, like picture recognition, and in six years has obtained shoppers like BASF, Ally, Moody’s, Bosch, and others.
Along with promoting its fashions on to main system producers, Multiverse provides its compressed fashions by way of an API hosted on AWS that any developer can use, usually at decrease token charges than opponents.
