Evidence-backed European AI register

Multiverse Computing

Reviewed identity, roles, model listings, pricing and evidence for Multiverse Computing.

Dataset compiled 2026-08-27 · evidence over inference

model labinference provider6 model listings10 token-price rows

Reviewed identity

Headquarters scope: EU headquarters
Headquarters record: San Sebastián (Donostia); additional office in Barcelona
Legal entity: S.L. (Spanish private limited company)
Ownership: VC-backed private
Founded: 2019
Confidence: high
Last verified: 2026-08-22

Links

Official website

Open interactive profile

Evidence boundaries

Organisation listings identify a reviewed canonical record. Model listings show an association in the source register; they do not by themselves establish model creation, current API availability, hosting location or infrastructure ownership.

Model listings

Hypernova-60B-2605catalogued with this organisation · origin not asserted
Type
text
Parameters
60B total MoE, 4.8B active
Context
131,072 tokens
Languages
unknown
Licence
Apache-2.0
Release
2026-02-26
Hypernova-60B-2602 (+ GGUF) / HyperNova-60Bcatalogued with this organisation · origin not asserted
Type
text
Parameters
~59–60B
Context
unknown
Languages
unknown
Licence
Apache-2.0 (family)
Release
2026-02
LittleLamb / LittleLamb-ToolCallingcatalogued with this organisation · origin not asserted
Type
text
Parameters
0.3B
Context
unknown
Languages
unknown
Licence
unknown
Release
2026-04-28
Pulsar-16B (BF16 / FP8 / NVFP4)catalogued with this organisation · origin not asserted
Type
text
Parameters
16B (NVFP4 variant listed at 9B)
Context
unknown
Languages
unknown
Licence
unknown
Release
unknown
Qwen3-Next-80B-A3B-Thinking-Uncensoredcatalogued with this organisation · origin not asserted
Type
text
Parameters
80B
Context
unknown
Languages
unknown
Licence
unknown
Release
unknown
Slim series (Llama 4 Scout Slim, Llama 3.3 70B Slim, Llama 3.1 8B Slim, Mistral Small 3.1 Slim, DeepSeek R1 Slim, Whisper Large V3 Turbo Slim)catalogued with this organisation · origin not asserted
Type
text / asr
Parameters
50–60% parameter reduction vs originals (Llama 4 Scout 50%, Llama 3.3 70B 50%, Llama 3.1 8B 60%, Mistral Small 3.1 50%)
Context
unknown
Languages
as per base models
Licence
inherits base model licences (Llama Community Licence, Apache-2.0 for Mistral/Whisper) — not independently verified per model
Release
2025–2026

Published token prices

Model	Input €/1M	Output €/1M	Currency	Evidence
Carina 60B	€0.0427	€0.0684	USD	price list
GLM 5.1	€0.812	€2.69	USD	price list
GLM 5.1 Uncensored	€0.8548	€2.56	USD	price list
GLM 5.2	€0.9403	€2.99	USD	price list
Hypernova 60B	€0.0342	€0.1197	USD	price list
Mistral Small 3.1	€0.094	€0.1453	USD	price list
Mistral Small 3.1 Slim	€0.0427	€0.0684	USD	price list
Nemotron 3 Nano Omni	€0.171	€0.6838	USD	price list
OpenAI gpt-oss-120b	€0.0427	€0.1966	USD	price list
Qwen 3.6 27B Uncensored	€0.1282	€0.7693	USD	price list

Profile note

Spain's first AI unicorn and the clearest Iberian example of a real, published per-token price list. Its business is quantum-inspired tensor-network compression (CompactifAI) of other people's models — so most of what it serves is compressed Llama / Mistral / DeepSeek / GLM / gpt-oss rather than in-house pretraining, though Hypernova and LittleLamb are its own distilled releases. 160+ patents; customers include Iberdrola, Bosch, Bank of Canada.