Mmbert Embed 32K 2d Matryoshka measured on 7 of 261 Embeddings benchmarks7
BenchmarkByScoreRank
No benchmark matches
price not verifiedprice
19.5MTEB Portuguese v1
Jan 2026released
768Dimensions
33K tokensMax input
Alternatives?
ModelScoreCostPriceWhat changes
BetterEmbeddingGemma 300MunstatedOpen+10%—$0.002Max input: 33K tokens → 2K tokensaOpenGiga Embeddings Instruct 3B 0826unstated+75%——Max input: 33K tokens → 8K tokens+1
BetterQwen 3 Embedding 0.6BunstatedOpen+16%—$0.01Memory (4-bit): 0.2 GB → 0.4 GBtBetterGte LargeunstatedOpen+13%—$0.01Max input: 33K tokens → 1K tokensBetterQwen 3 Embedding 8BunstatedOpen+13%—$0.01Memory (4-bit): 0.2 GB → 4.5 GBModelScoreCostPriceWhat changes
Nothing here yet.
ModelScoreCostPriceWhat changes
EmbeddingGemma 300MunstatedOpen+10%—$0.002Max input: 33K tokens → 2K tokensQwen 3 Embedding 0.6BunstatedOpen+16%—$0.01Memory (4-bit): 0.2 GB → 0.4 GBtGte LargeunstatedOpen+13%—$0.01Max input: 33K tokens → 1K tokensQwen 3 Embedding 8BunstatedOpen+13%—$0.01Memory (4-bit): 0.2 GB → 4.5 GB
Text Embedding 3 Smallunstated+13%—$0.02Max input: 33K tokens → 8K tokensModelScoreCostPriceWhat changes
aGiga Embeddings Instruct 3B 0826unstatedOpen+75%——Max input: 33K tokens → 8K tokens+1aGiga Embeddings Instruct 10B A1.8B 0826unstatedOpen+74%——Max input: 33K tokens → 8K tokens+1
KaLM Embedding Gemma3 12B 2511Open+62%——Memory (4-bit): 0.2 GB → 7.1 GBBBidirLM 1.7B EmbeddingOpen+49%——Max input: 33K tokens → 1K tokens+1
Llama Embed Nemotron 8BOpen+49%——Memory (4-bit): 0.2 GB → 4.5 GBRun it yourself
estimateFits on✓ 8 GB (phone, laptop) · full quality✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceApache-2.0free for commercial use
Benchmarks
in Embeddings & RetrievalFAQ
How much does Mmbert Embed 32K 2d Matryoshka cost?›
There is no per-unit API price: price not verified.
How good is Mmbert Embed 32K 2d Matryoshka, and what is it best at?›
#94 of 191 in Embeddings.
Is Mmbert Embed 32K 2d Matryoshka better than Snowflake Arctic Embed M?›
On MTEB Portuguese v1 · Multilabel Classification · Portuguese, Mmbert Embed 32K 2d Matryoshka scores 19.5 vs 19.5 — about the same.
When was Mmbert Embed 32K 2d Matryoshka released? Is there a newer version?›
Released 2026-01-26. It is the current version in its line.
Is Mmbert Embed 32K 2d Matryoshka open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Mmbert Embed 32K 2d Matryoshka locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
What is Mmbert Embed 32K 2d Matryoshka?›
Mmbert Embed 32K 2d Matryoshka is an open-weights embeddings & retrieval model from llm-semantic-router, released Jan 2026.
What are the best alternatives to Mmbert Embed 32K 2d Matryoshka?›
EmbeddingGemma 300M (better), Giga Embeddings Instruct 3B 0826 (open weights).