E5-Mistral
Mistral-7B turned into a state-of-the-art text embedding model.
About
E5-Mistral-7B-instruct, released in January 2024, fine-tunes Mistral-7B-v0.1 on a mixture of synthetic and labeled data to produce text embeddings. It was the first LLM-decoder-based embedding model to top the MTEB leaderboard, showing that large decoder models could beat much smaller specialized encoders at retrieval, clustering, and classification.
Its success spawned a wave of LLM-based embedding models and retrieval stacks. It is released under the MIT license, though the model card notes the Mistral base was trained mainly on English, so English use is recommended.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 82; Trending score 76/100 from researcher assessment.