# NVIDIA releases TensorRT‑LLM 1.4.0rc1 with KV cache manager V2

> The new release adds support for several models, makes the KV cache manager V2 the default, and removes unused request fields.

Oossa · 2026-10-22 · https://oossa.com/en/nvidia-releases-tensorrt-llm-1-4-0rc1-with-kv-cache-manager-v2

NVIDIA posted TensorRT‑LLM version 1.4.0rc1 on October 22 2026. The update makes the KV cache manager V2 the default API, a breaking change for developers, and drops the now‑unused LlmRequest fields. It also adds model support for DeepSeek‑V4.1‑Flash, NVFP4 GLM‑5.3‑Flash, and Qwen3.8‑Flash‑Next, among others.

## The facts

- Release date: 2026‑10‑22 ("Thu Oct 22 2026")
- KV cache manager V2 is now the default; LlmRequest fields removed

## Why it matters

Developers using TensorRT‑LLM must update their code to the new KV cache API, or their applications will fail to start.

## Sources & references

1. [NVIDIA/TensorRT-LLM v1.4.0rc1](https://github.com/NVIDIA/TensorRT-LLM/releases/tag/v1.4.0rc1) – TensorRT-LLM, 2026-10-22

Last updated: 2026-10-22