# EmbeddingGemma

27 Hacker News commenters recommended EmbeddingGemma and 4 criticized it, November 2025 to November 2026. It was recommended in 3 of 13 months.

**Verdict:** Open embedding model people praise for quality, low RAM use and its Apache 2.0 license. Critics note weaker accuracy than cloud models for speech and some tasks.

## Why people recommend it

- Strong benchmarks and quality (4 comments): "Better quality, less RAM usage, faster embeddings, I'm fucking pumped for it." ([source](https://news.ycombinator.com/item?id=49992946)) "has incredible benchmarks and a 2k context window" ([source](https://news.ycombinator.com/item?id=46081800)) "which has incredible benchmarks" ([source](https://news.ycombinator.com/item?id=46086119))
- Small and efficient for its quality (3 comments): "270M for text only is great compared to older embedding models" ([source](https://news.ycombinator.com/item?id=49980771)) "embeddiggemma because it's crazy efficient" ([source](https://news.ycombinator.com/item?id=48017877)) "Better quality, less RAM usage, faster embeddings, I'm fucking pumped for it." ([source](https://news.ycombinator.com/item?id=49992946))
- Open weights under Apache 2.0 (2 comments): "Hats off to google for offering OSS (or at least open weights + license) a model" ([source](https://news.ycombinator.com/item?id=49984069)) "I really appreciate that EmbeddingGemma 2 is under the Apache 2.0 license." ([source](https://news.ycombinator.com/item?id=49983751))
- Multimodal text, image and video use (2 comments): "And if your indexing software uses it correctly you can also do video deposition searches" ([source](https://news.ycombinator.com/item?id=49994210)) "It's also very neat that this can be used for "Jev"-like tasks with text and image." ([source](https://news.ycombinator.com/item?id=49984406))

## What people warn about

- Transcription accuracy below cloud models (3 comments): "WER is not good enough for professional use, fine for hobby projects" ([source](https://news.ycombinator.com/item?id=50070369)) "The accuracy is good, but not in the range of cloud models" ([source](https://news.ycombinator.com/item?id=50071209)) "the overall transcription quality is very low" ([source](https://news.ycombinator.com/item?id=50073142))

## Alternatives hackers recommend

- [Ternlight](https://hackerapproved.com/p/ternlight.md): 17 recommend
- [CLIP by OpenAI](https://hackerapproved.com/p/clip.md): 10 recommend
- [Gemini Embedding 2](https://hackerapproved.com/p/gemini-embedding-2.md): 9 recommend
- [Qwen3-Embedding](https://hackerapproved.com/p/qwen3-embedding-0-6b.md): 8 recommend
- [ModernBERT](https://hackerapproved.com/p/modernbert.md): 7 recommend

## Cite as

Hacker Approved, "EmbeddingGemma", https://hackerapproved.com/p/embeddinggemma-2/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).