AITOPAI热报
ModelsMulti-sourceSelected

vLLM adds Day-0 support for Recursive's Flint model

官方账号vLLM@vllm_project
Post· 8 sources
View the post
Why we picked it

A sparse 2.5B-active architecture hitting 500 TPS on one GB200 shows how far inference efficiency now goes without speculative decoding, a benchmark point worth tracking against dense models of similar quality. Teams serving long-context workloads can run Flint in vLLM immediately in either NVFP4 or BF16. Builders evaluating open-weight model serving stacks should compare its throughput on their own GPUs.

vLLM announced Day-0 support for Flint, a 40B-parameter model from Recursive (Recursive_SI) with only 2.5B active parameters per token. vLLM's custom NVFP4 kernels reach 500 tokens per second on a single NVIDIA GB200, even at 100K context without speculative decoding. Both NVFP4 and BF16 checkpoints are runnable in vLLM today, 25 Sep 2025.

Original · vLLM

Day-0 support for @Recursive_SI's Flint is here in vLLM! 🚀

40B parameters, just 2.5B active per token. Thanks to the small size, our custom NVFP4 kernels push it to 500 TPS on a single GB200, even at 100K context without speculative decoding.

Both NVFP4 and BF16 are ready to run today! https://t.co/KzlqAEqoHM

#vLLM#Flint#NVFP4#open-source models#inference

Keep reading

Recent stories with the same tag or section
ProductsBuildersMulti-sourceSelected04:07
PyTorchCon Demo Theater: 13 Sessions on Training, Inference and Agents

The lineup reveals where the PyTorch ecosystem is concentrating effort: inference speed for video models, MoE expert parallelism on NVL72, and fault-tolerant training that survives link and node failures live. Engineers choosing accelerator backends can use these demos to compare AWS Trainium, Intel AI PCs, TPUs and Rebellions hardware on real workloads. The Huawei CRCR CI relay is also a practical pattern for teams maintaining forks against upstream PyTorch.

#PyTorch#PyTorchCon#Wan 2.2#vLLM
Event
官方账号PyTorch@PyTorch11 sourcesSource
ProductsBuildersSelected02:54
vLLM-TorchTPU brings native PyTorch serving to Google TPUs

The practical effect is that vLLM's optimizations — prefix caching and speculative decoding — now apply to TPU hardware, giving teams a second accelerator option beyond NVIDIA GPUs for self-hosted inference. Organisations already locked into the PyTorch and vLLM stack can evaluate TPU economics without rewriting their serving code. It also signals Google's continued push to open TPU inference beyond JAX.

#vLLM#TPU#PyTorch#Google
官方账号vLLM@vllm_projectSource
ProductsBuildersSelected02:01
FlyDSL kernel DSL now supports PyTorch on AMD MI350 GPUs

AMD MI350 support matters because GPU kernel tooling has historically centred on NVIDIA's stack, and this gives teams serving open models a Python-native path to tune kernels without CUDA-only tools. Benchmark and deployment engineers running vLLM on AMD accelerators can adopt the DSL directly from PyTorch's ecosystem and measure the 3.01x serving gain against their own baselines.

#FlyDSL#PyTorch#AMD#vLLM
官方账号PyTorch@PyTorchSource
IndustryBuildersSelected01:07
Cohere's Rachad Alao on open science for foundation models

The talk links open model releases to concrete security practices like independent auditing and dynamic safeguards, rather than treating openness and safety as opposing goals. Researchers and ML engineers can follow the named stack — PyTorch, TorchTitan and vLLM — to reproduce or audit open models themselves. Cohere's stance is notable as a US-listed enterprise lab that publicly backs open science. origin

#Cohere#PyTorch#TorchTitan#vLLM
官方账号PyTorch@PyTorchSource