Skip to main content
Fordel Studios logoFordel Studios
Start a Conversation
Back to Pulse
TechCrunch

Spanish ‘soonicorn’ Multiverse Computing releases free compressed AI model

Read the full articleSpanish ‘soonicorn’ Multiverse Computing releases free compressed AI model on TechCrunch
↗

What Happened

Spanish startup Multiverse Computing has released a new version of its HyperNova 60B model on Hugging Face that, it says, bests Mistral's model.

Our Take

'Beats Mistral'—on what? Inference speed? Accuracy? File size? Multiverse isn't saying. Releasing free compressed models is smart marketing, but 'best' means nothing without benchmarks.

The real test: does it run on edge hardware without tanking quality? Third-party evals on standard benchmarks are the only thing that matters. Until then, I'm reading this as 'we compressed our model and it didn't explode.'

What To Do

Find their actual benchmark claims—if they're vague, skip it.

Cited By

React

Loading comments...