GTE-Qwen2 7B
Alibaba's 7B LLM-based embedder — Apache 2.0 with 32K context for long-document retrieval.
About
GTE-Qwen2-7B-Instruct (2024, Alibaba's GTE team) applies the LLM-embedder recipe to Qwen2-7B: instruction-tuned embeddings with 32K context for long-document retrieval.
Apache 2.0 licensed — commercial use allowed. 103k Hugging Face downloads. The 32K context is the standout trait — few embedders handle truly long documents. Like E5-Mistral, it's a 7B model that's GPU-hungry for embedding workloads, which limited adoption. A 2024 model in a field that has since moved to more efficient architectures, but the long-context capability remains distinctive.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-claimed (Alibaba): strong MTEB performance — treat as provider-claimed. Independent standardized scores are limited.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 6/10
Scored 6/10 (metrics tier). Apache 2.0 with a genuinely distinctive 32K context for long-document retrieval. Docked four points: (1) 7B is GPU-hungry for embedding work; (2) modest adoption (103k downloads); (3) benchmark claims are provider-reported; (4) 2024-era with more efficient alternatives now. For long-document retrieval on capable hardware.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 4.0/10
- Openness · license + docs
- 9.1/10
- Manual review · sanity read + red-flag check
- 6.0/10
- · HF downloads 102,602 (peer-normalized)
- · released 850 days ago
- · manual review: 32K context is distinctive; 7B GPU-hungry; modest adoption; license: Apache-2.0