Pearl Serverless Inference
Pay-per-token OpenAI-compatible endpoints for open models on Pearl's optimized serving stack
Community-submitted · content unverified. This organization has not verified ownership or the information in this profile.
About
Pearl Serverless Inference provides OpenAI-compatible API endpoints, billed per token, for leading open models including DeepSeek, Qwen, Gemma and GLM variants with context windows up to 1M tokens. It runs on Pearl Research Labs' optimized serving stack, which uses techniques such as KV-cache quantization (PearlQuant), custom MoE kernels and speculative decoding to deliver more tokens per second at lower cost.
- Website
- pearlresearch.ai/pricing
- Added via
- web
- Ownership
- unclaimed
Where to go
For people
- Homepagepearlresearch.ai/pricing
For agents
Nothing listed yet.
Product details
- Product kind
- Api