EmbeddingGemma 2
My comment on EmbeddingGemma 2 — Hacker News. I really appreciate that EmbeddingGemma 2 is under the Apache 2.0 license. For embedding models in particular, I don't think it makes sense to use a closed, proprietary, hosted-only model. Most applications of embedding models involve calculating thousands or even millions of embedding vectors and storing them for later comparison. If your model is…
The Apache 2.0 license for EmbeddingGemma 2 is appreciated as it allows for the embedding of models. Most applications of embedding models require calculating thousands or millions of embedding vectors which need to be stored for later comparison. If the model is proprietary, the vendor may stop offering it in the future and release a better model, but users would still need to pay to re-calculate the stored vectors.
In April 2024, OpenAI offered to cover the financial cost of users re-embedding content with new models, but this cannot be relied upon by every provider. Additionally, hosting the model oneself is not preferred. Users would rather pay a provider for a hosted model, knowing that if they ever stop hosting it, the open weights version can be run by the user or another vendor. This perspective was shared in a beat by Simon Willison on October 6th, 2026.
Written by urgent.news from Simon Willison's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- EmbeddingGemma 2 blog.google
