Wispr Flow - Don’t type, just speak
Mini tools| Model | Input | Cached input | Output | Unit |
|---|---|---|---|---|
|
pplx embed v1 0.6b
Perplexity
|
$0.004 | - | $0 | per 1M tokens |
|
pplx embed v1 4b
This model
Perplexity
|
$0.03 | - | $0 | per 1M tokens |
|
pplx embed context v1 0.6b
Perplexity
|
$0.008 | - | $0 | per 1M tokens |
|
pplx embed context v1 4B
Perplexity
|
$0.05 | - | $0 | per 1M tokens |
pplx embed v1 4b
By Perplexity
pplx-embed-v1-4b is a 4-billion-parameter dense text embedding model built on a Qwen3 backbone continued-pretrained with diffusion. It produces 2560-dimensional embeddings natively quantized to INT8 or binary precision, enabling efficient large-scale retrieval. Embeddings are unnormalized and should be compared with cosine similarity. The model supports a 32K token context window and Matryoshka Representation Learning (MRL) for flexible output dimensionality reduction. Mean pooling is used. It is instruction-aware, allowing direct embedding without instruction-prefix tuning for most use cases. Multilingual support is included. Deployable via the Perplexity API, SentenceTransformers, ONNX Runtime, and Text Embeddings Inference (v1.9.2+). Released under the MIT license.
Text
Gen 7
Overview
A 4B-parameter text embedding model built on diffusion-pretrained Qwen3. Produces 2560-dimensional dense embeddings with INT8 and binary quantization, 32K context window, and Matryoshka Representation Learning (MRL) support. Designed for semantic search and web-scale retrieval. Multilingual. Available via API, SentenceTransformers, ONNX, and Text Embeddings Inference.
Pricing
Compare pplx embed v1 4b with other models listed in the same vendor pricing tiers and context lengths.
Embeddings
About Perplexity
Perplexity is a technology company that specializes in artificial intelligence and machine learning solutions.
Industry:
Artificial Intelligence
Company Size:
247
Location:
San Francisco, California, US
Website:
perplexity.ai
Last updated: June 25, 2026
