---
format: "aidr-story-markdown/v1"
id: "5a577b50ff19fb234b5e4ae609cf043eb7c39f6b2b860de8163bfd9541da706a"
canonical_url: "https://aidr.today/5a577b50?lang=en"
title: "Perplexity and Baseten Host DeepSeek V4 Pro in US, Cutting Frontier Costs 62%"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-08-19T02:48:10.000Z"
category: "Infra"
topics: ["deepseek","inference","llm-pricing"]
source_urls: ["https://huggingnews.com/ai/perplexity-and-baseten-host-deepseek-v4-pro-in-us-cutting-frontier-costs-318ec25a"]
summary: "The 0813 version of DeepSeek V4 Pro reached 147 tokens per second in tests by Artificial Analysis, the highest speed among currently available inference provid…"
---

# Perplexity and Baseten Host DeepSeek V4 Pro in US, Cutting Frontier Costs 62%

> [Open the canonical story](<https://aidr.today/5a577b50?lang=en>)

**Published:** 2026-08-19T02:48:10.000Z
**Category:** Infra
**Topics:** deepseek, inference, llm\-pricing

## Summary

The 0813 version of DeepSeek V4 Pro reached 147 tokens per second in tests by Artificial Analysis, the highest speed among currently available inference provid…

## Sources

- [Story source](<https://huggingnews.com/ai/perplexity-and-baseten-host-deepseek-v4-pro-in-us-cutting-frontier-costs-318ec25a>)