---
format: "aidr-story-markdown/v1"
id: "6fe126280bd967f73db93cbee6c58c62a710cf5cc43c9330c4ec7781389d431b"
canonical_url: "https://aidr.today/6fe12628?lang=en"
title: "llama.cpp Releases v0.9.0 With Support for Native FP8 and Multi Node Tensor Parallelism"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-04T20:37:18.000Z"
category: "Open Source"
topics: ["llm","open-source","framework"]
source_urls: ["https://marketbrief.now/ai/llamacpp-releases-v090-with-support-for-native-fp8-and-multi-node-tensor-ec738073","https://huggingnews.com/ai/llamacpp-releases-v090-with-support-for-native-fp8-and-multi-node-tensor-ec738073"]
summary: "llama.cpp releases v0.9.0 with support for native FP8 and Multi Node Tensor Parallelism. This update enables tensor parallels across multiple nodes via NCCL. The version also adds speech capabilities incorporating Chatterbox TTS and Nemotron 3.5 Transcribe."
---

# llama\.cpp Releases v0\.9\.0 With Support for Native FP8 and Multi Node Tensor Parallelism

> [Open the canonical story](<https://aidr.today/6fe12628?lang=en>)

**Published:** 2026-11-04T20:37:18.000Z
**Category:** Open Source
**Topics:** llm, open\-source, framework

## Summary

llama\.cpp releases v0\.9\.0 with support for native FP8 and Multi Node Tensor Parallelism\. This update enables tensor parallels across multiple nodes via NCCL\. The version also adds speech capabilities incorporating Chatterbox TTS and Nemotron 3\.5 Transcribe\.

## Sources

- [Story source](<https://marketbrief.now/ai/llamacpp-releases-v090-with-support-for-native-fp8-and-multi-node-tensor-ec738073>)
- [Story source](<https://huggingnews.com/ai/llamacpp-releases-v090-with-support-for-native-fp8-and-multi-node-tensor-ec738073>)