multilingual-e5-base
Will it run on your hardware?
Pick your GPU memory - see which quantizations fit, and the cheapest card for the rest
Need an exact figure for your context length? Use the VRAM calculator.
Run it locally
Copy-paste - running in under a minute
vllm serve intfloat/multilingual-e5-baseNew to this? Start with Ollama · serve to many users with vLLM.
Deep dive
Notes, sources, and the full write-up
multilingual-e5-base is an open-weight text embedding model published by intfloat on Hugging Face in May 2023, with 280M parameters under the mit license.
Among the 19 text embedding models in this directory it ranks 12th by monthly Hugging Face downloads, and 9 of the others are larger.
HardwareLink to this section
Unquantized FP16 weights take about 1 GB; 8-bit brings that to ~1 GB and 4-bit (Q4_K_M) to ~1 GB, before activations and any context or image buffers. At this size it runs on almost any recent GPU, on Apple Silicon, and usable on a modern CPU. Check exact numbers for your card with the VRAM calculator.
Similar text embedding modelsLink to this section
- nomic-embed-text-v1.5 - 140M, 11,463,751 downloads/month
- multilingual-e5-large - 560M, 7,318,858 downloads/month
- bge-large-en-v1.5 - 340M, 6,814,594 downloads/month
- harrier-oss-v1-0.6b - 600M, 234,641 downloads/month
- voyage-4-nano - 350M, 156,246 downloads/month
PopularityLink to this section
multilingual-e5-base was downloaded 4,651,836 times on Hugging Face in the last month (76,691,408 all-time) and has 422 likes.
Run it locallyLink to this section
Weights and usage instructions are on the intfloat/multilingual-e5-base model card; it loads with sentence-transformers.
Local runtimes that fit this model:
- transformers - the reference Python library most model cards target
PaperLink to this section
Frequently asked
Quick answers to common questions
How much VRAM does multilingual-e5-base need?
multilingual-e5-base with 280M parameters needs approximately 1 GB at Q4_K_M quantization. Use our VRAM calculator for an exact estimate.
Is multilingual-e5-base better than other intfloat models?
multilingual-e5-base has 280M parameters - a strong choice for embeddings.
What license is multilingual-e5-base under?
multilingual-e5-base is released under the mit license, making it suitable for most commercial and personal projects.
What hardware runs multilingual-e5-base well?
With 280M parameters, multilingual-e5-base is small enough for most consumer GPUs, Apple Silicon Macs, and many CPUs. Check our hardware directory for specific recommendations.
What is the best quantization for multilingual-e5-base?
Q4_K_M is the recommended sweet spot - ~98% of FP16 quality at ~27% of the size. Q5_K_M (~1 GB) is an option if you have spare VRAM. Use our VRAM calculator to compare.
What models compete with multilingual-e5-base?
multilingual-e5-base competes with other models in its class. Browse our model directory for comparisons, benchmarks, and community reviews to find the best fit.
Nearby options
Similar models and compatible hardware by spec