---
format: "aidr-story-markdown/v1"
id: "b99beecb4669479f86b86323705160881d050bd4fc54a1e5599f3be4487a5375"
canonical_url: "https://aidr.today/b99beecb?lang=en"
title: "Google's EmbeddingGemma 2 Rivals OpenAI on Consumer CPUs"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-14T09:09:17.000Z"
category: "Data"
topics: ["google","llm","data-engineering"]
source_urls: ["https://huggingnews.com/ai/googles-embeddinggemma-2-rivals-openai-on-consumer-cpus-eb76065a","https://x.com/dorkitude/status/2109811136334041468","https://x.com/tomaarsen/status/2109974446094295248","https://x.com/LearnOpenCV/status/2110291437174567257","https://x.com/tomaarsen/status/2110126159367479684","https://x.com/TheTuringPost/status/2109817905177378908","https://x.com/googlegemma/status/2110044339351191707","https://marketbrief.now/ai/googles-embeddinggemma-2-rivals-openai-on-consumer-cpus-eb76065a"]
summary: "Google released EmbeddingGemma 2, a 740M parameter multimodal model that benchmarks comparably to OpenAI's embedding-3-large and Qwen 8B while running on consumer CPUs rather than GPUs. The model has become the 4th most liked in the company's history after surpassing EmbeddingGemma 1 in popularity. Developers have used the model to build DigUp, an open-source tool for Mac that indexes local files and podcasts, with Google employing Matryoshka embeddings to store only the first 256 of 768 values in a vector, reducing storage needs to 1/3 while maintaining about 95% of the retrieval quality for visuals and speech."
---

# Google's EmbeddingGemma 2 Rivals OpenAI on Consumer CPUs

> [Open the canonical story](<https://aidr.today/b99beecb?lang=en>)

**Published:** 2026-10-14T09:09:17.000Z
**Category:** Data
**Topics:** google, llm, data\-engineering

## Summary

Google released EmbeddingGemma 2, a 740M parameter multimodal model that benchmarks comparably to OpenAI's embedding\-3\-large and Qwen 8B while running on consumer CPUs rather than GPUs\. The model has become the 4th most liked in the company's history after surpassing EmbeddingGemma 1 in popularity\. Developers have used the model to build DigUp, an open\-source tool for Mac that indexes local files and podcasts, with Google employing Matryoshka embeddings to store only the first 256 of 768 values in a vector, reducing storage needs to 1/3 while maintaining about 95% of the retrieval quality for visuals and speech\.

## Sources

- [Story source](<https://huggingnews.com/ai/googles-embeddinggemma-2-rivals-openai-on-consumer-cpus-eb76065a>)
- [Story source](<https://x.com/dorkitude/status/2109811136334041468>)
- [Story source](<https://x.com/tomaarsen/status/2109974446094295248>)
- [Supporting source](<https://x.com/LearnOpenCV/status/2110291437174567257>)
- [Supporting source](<https://x.com/tomaarsen/status/2110126159367479684>)
- [Supporting source](<https://x.com/TheTuringPost/status/2109817905177378908>)
- [Supporting source](<https://x.com/googlegemma/status/2110044339351191707>)
- [Story source](<https://marketbrief.now/ai/googles-embeddinggemma-2-rivals-openai-on-consumer-cpus-eb76065a>)