// Jina embeddings & rerankers · Embedding
jina-embeddings-v4: VRAM & local deployment
3.75B parameters. About 3.0 GB at 4-bit and 8.8 GB at BF16.
released 2025-05-07licence: Licence not statedmemory figures are estimates
01Specs
| Publisher | Jina AI · jinaai |
|---|---|
| Family / series | Jina embeddings & rerankers · jina-embeddings |
| Type | Embedding (visual-document-retrieval) |
| Parameters | 3.75B Source: counted from official safetensors weights |
| Released weights | BF16 · 7.5 GBdownload |
| Licence | · Licence not statedThe model card metadata does not state a licence; check the repository files. |
| Released | 2025-05-07 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Embedding models are small; memory grows with batch size and sequence length.
| Precision | Weights | Weights + overhead | Smallest GPU | Mac |
|---|---|---|---|---|
| BF16 / FP16 (original) | 7.5 GB | 8.8 GB | 12GB GPU | Mac 16GB |
| FP8 / INT8 / Q8_0 | 4.0 GB | 4.9 GB | 8GB GPU | Mac 16GB |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 2.3 GB | 3.0 GB | 8GB GPU | Mac 16GB |
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- vLLM — mentioned in the official model card
- Transformers — mentioned in the official model card
- Sentence Transformers — mentioned in the official model card
- llama.cpp — official GGUF weights
- MLX — official MLX weights
04Official repositories & quantised builds
5 official repo(s) from jinaai.
05Other Jina embeddings & rerankers models
All 16 →| Model | Params | 4-bit total | Released |
|---|---|---|---|
| jina-embeddings-v3 · same series | 572M | 0.9 GB | 2024-09-05 |
| jina-reranker-v3.5 | 597M | 0.9 GB | 2026-07-14 |
| jina-embeddings-v5-omni-nano | 986M | 1.2 GB | 2026-04-01 |
| jina-embeddings-v5-omni-small | 1.63B | 1.6 GB | 2026-04-01 |
| jina-embeddings-v5-omni-nano-retrieval | 948M | 1.1 GB | 2026-03-31 |
| jina-embeddings-v5-omni-small-retrieval | 1.56B | 1.5 GB | 2026-03-31 |
| jina-embeddings-v5-text-nano | 212M | 0.6 GB | 2026-01-22 |
| jina-embeddings-v5-text-small | 596M | 0.9 GB | 2026-01-22 |
| jina-reranker-v3 | 597M | 0.9 GB | 2025-09-18 |
06Sources
Data checked 2026-04-08. Parameters, context and licence are copied from the official repository; memory figures are our estimates.