pplx-embed-v1-0.6b

pplx-embed-v1 and pplx-embed-context-v1 are state-of-the-art text embedding models optimized for real-world, web-scale retrieval tasks.

• Use pplx-embed-v1 for independent text embedding (queries, documents, semantic search)
• Use pplx-embed-context-v1 for document chunks in RAG systems where surrounding context matters

[!IMPORTANT] pplx-embed-v1 and pplx-embed-context-v1 natively produce unnormalized int8-quantized embeddings. Ensure that you compare them via cosine similarity.

Models

All models are built on diffusion continued pre-trained Qwen3 at Perplexity AI.

Many modern embedding models rely on instruction tuning, where users prepend an instruction string to the text being embedded. This can yield a 2%-3% lift on benchmarks, but it also introduces prompt-selection overhead and can make indexing pipelines brittle (small instruction changes can shift embedding space). We deliberately avoid this requirement: you can embed the text you want to index directly, without having to choose or maintain an instruction prefix.

Usage

curl -X POST https://api.perplexity.ai/v1/embeddings \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": [
      "Scientists explore the universe driven by curiosity.",
      "Children learn through curious exploration.",
      "Historical discoveries began with curious questions.",
      "Animals use curiosity to adapt and survive.",
      "Philosophy examines the nature of curiosity."
    ],
    "model": "pplx-embed-v1-0.6b"
  }'

from sentence_transformers import SentenceTransformer

model = SentenceTransformer(
    "perplexity-ai/pplx-embed-v1-0.6B",
    trust_remote_code=True
)

texts = [
    "Scientists explore the universe driven by curiosity.",
    "Children learn through curious exploration.",
    "Historical discoveries began with curious questions.",
    "Animals use curiosity to adapt and survive.",
    "Philosophy examines the nature of curiosity.",
]

embeddings = model.encode(texts) # Shape: (5, 1024), quantized to int8
embeddings = model.encode(texts, quantization="binary") # Shape: (5, 1024), quantized to binary

import onnxruntime as ort
from transformers import AutoTokenizer
import numpy as np

tokenizer = AutoTokenizer.from_pretrained("perplexity-ai/pplx-embed-v1-0.6b", trust_remote_code=True)
session = ort.InferenceSession("onnx/model.onnx")


texts = [
    "Scientists explore the universe driven by curiosity.",
    "Children learn through curious exploration.",
    "Historical discoveries began with curious questions.",
    "Animals use curiosity to adapt and survive.",
    "Philosophy examines the nature of curiosity.",
]

tokenized = tokenizer(
    texts,
    padding=True,
    truncation=True,
    return_tensors="np"
)

onnx_inputs = {
    "input_ids": tokenized["input_ids"].astype(np.int64),
    "attention_mask": tokenized["attention_mask"].astype(np.int64),
}

# Run inference
onnx_embeddings = session.run([out.name for out in session.get_outputs()], onnx_inputs)

# ONNX produces both int8 and binary precision embeddings:
int8_embeddings = onnx_embeddings[2]
binary_embeddings = onnx_embeddings[3]
packed_embeddings = np.packbits(binary_embeddings != -1, axis=-1)

[!NOTE] Text Embeddings Inference v1.9.2+ is required.

[!IMPORTANT] Currently, only int8-quantized embeddings are available via TEI. Remember to use cosine similarity with unnormalized int8 embeddings.

• CPU w/ Candle:

docker run -p 8080:80 ghcr.io/huggingface/text-embeddings-inference:cpu-1.9 --model-id perplexity-ai/pplx-embed-v1-0.6B --dtype float32

• CPU w/ ORT (ONNX Runtime):

docker run -p 8080:80 ghcr.io/huggingface/text-embeddings-inference:cpu-1.9 --model-id onnx-community/pplx-embed-v1-0.6B --dtype float32

• GPU w/ CUDA:

docker run --gpus all --shm-size 1g -p 8080:80 ghcr.io/huggingface/text-embeddings-inference:cuda-1.9 --model-id perplexity-ai/pplx-embed-v1-0.6B --dtype float32

If you hit OOM during warmup, lower --max-batch-tokens and --max-client-batch-size. Set --max-batch-tokens to max_sequence_length × batch_size (e.g., 2048 tokens × 8 sequences = 16384).

Alternatively, when running in CUDA you can use the architecture / compute capability specific container instead of the cuda-1.9, as that includes the binaries for Turing, Ampere, Hopper and Blackwell, so using a dedicated container will be lighter e.g., ampere-1.9.

And then you can send requests to it via cURL to /embed:

curl http://0.0.0.0:8080/embed \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": [
      "Scientists explore the universe driven by curiosity.",
      "Children learn through curious exploration.",
      "Historical discoveries began with curious questions.",
      "Animals use curiosity to adapt and survive.",
      "Philosophy examines the nature of curiosity."
    ],
    "normalize": false
  }'

Technical Details

For comprehensive technical details and evaluation results, see our paper on arXiv: https://arxiv.org/abs/2602.11151.