News
ReleasesOVHcloud unveils Olala mini, its first in-house LLMChinaA 9B model beats a 1T model on FinQA, trained with Crusoe Managed RLReleasesThe internal metaphysics of an LLMReleasesMapping Language Model Internal Representations to Physical Laws and Scientific ReasoningResearchCan China keep rogue AI in check? Meet the start-up testing the top modelsFrontierMicrosoft aims to challenge OpenAI, Anthropic with cheaper, specialized AI modelsOpen modelsMicrosoft rolls out new first-party models and updates to agentic AI continuous improvementFrontierMicrosoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAIFrontierGoogle is making it easier to opt out of Gemini AI model training – here’s howFrontierOpenAI Unveils New Astra AI Model After Scrapping Prior ReleaseReleasesOVHcloud unveils Olala mini, its first in-house LLMChinaA 9B model beats a 1T model on FinQA, trained with Crusoe Managed RLReleasesThe internal metaphysics of an LLMReleasesMapping Language Model Internal Representations to Physical Laws and Scientific ReasoningResearchCan China keep rogue AI in check? Meet the start-up testing the top modelsFrontierMicrosoft aims to challenge OpenAI, Anthropic with cheaper, specialized AI modelsOpen modelsMicrosoft rolls out new first-party models and updates to agentic AI continuous improvementFrontierMicrosoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAIFrontierGoogle is making it easier to opt out of Gemini AI model training – here’s howFrontierOpenAI Unveils New Astra AI Model After Scrapping Prior Release
EveryLLM

National model programme, Finland

Helsinki

Silo AI and the University of Turku built the Poro and Viking models for Nordic languages on LUMI.

Helsinki is home to Silo AI, once Europe's largest private AI lab, which AMD bought in a deal announced in July 2024 for about US$665 million. With the University of Turku's TurkuNLP group, Silo AI trained open models for Finnish and the other Nordic languages on LUMI, a European supercomputer in Kajaani, Finland.

Poro and Viking

Poro, a 34-billion-parameter model for Finnish, English and code, was released in 2024 after training on LUMI, one of Europe's most powerful supercomputers, which runs on AMD graphics chips. Viking extended the work to Danish, Norwegian, Swedish and Icelandic. Finnish has only about five million speakers, so the models mixed in English and code to learn general skills.

Why Finland

Finland combines strong universities, a tradition of open-source software and a large public supercomputer run by its national IT centre for science, CSC. The AMD purchase made Silo AI's team part of a US chip maker's effort to compete with Nvidia, while its Nordic models remain openly available. Finland also hosts data centres drawn by its cool climate and clean power.

The weekly digest

Every Monday: the model of the day, new pages on EveryLLM, and the conferences and anniversaries of the week ahead.