Skip to main content

Ctrl+K

Get Started

Installation — CUDA
Installation — Apple Silicon
Installation — NPU
Installation — Intel XPU
Installation — Intel CPU
Installation — MUSA
Troubleshooting

Cookbook

Qwen
OpenMOSS
OpenAI
- Whisper ASR
NVIDIA
Fish Audio
- Fish Audio S2-Pro
- Fish Audio Drama 3
FunAudioLLM
- Fun-CosyVoice3
- Fun-ASR-Nano
OpenBMB
- MiniCPM-o
Boson AI
- Higgs TTS
MiniMax
- MiniMax Music 3
InclusionAI
Tencent
- AuK
- Hy-SimulMT
Zyphra
- ZONOS2
AutoArk AI
- ARK-ASR-3B
Mistral AI
- Voxtral TTS
Dots Studio
- dots.tts

General Usage

Omni Model Usage
Native Chat Stage
Audio Translations
TTS Model Usage
Process Topology, Replicas, and GPU Sharing
TTS Process Topology
Process Topology Migration
SGLang-Omni Router
Same-GPU Data Parallelism with CUDA MPS
Native Media Routes

MLX

MLX Path

Developer Reference

Architecture
API Server Design
Pipeline Overview
Config
Adding a Parameter
Communication
Reference Encode Service
Request-level profiler
Qwen3-ASR High-Concurrency Benchmark and Bottleneck Profile
Full-Duplex Guide
Duplex Load Generator Pacing
TTS benchmark runbook
TTS Model Integration
ASR benchmark runbook
MiniCPM-o offline benchmark runbook
RL Admin Control
CI Workflows and Labels
Bumping the SGLang Pin

Repository
Show source
Suggest edit
Open issue

.md

InclusionAI

InclusionAI#

Ming and LLaDA models from inclusionAI (StepFun).

Ming-Omni
LLaDA2.0-Uni
Ming-Omni-TTS
Ming-3.0-asr-tiny
Ming-3.0-audio-flash
Ming-3.0-TTS

previous

MiniMax Music 3

next

Ming-Omni

By SGLang-Omni Team

© Copyright 2025-2026, SGLang-Omni.

Last updated on Nov 02, 2026.