SGLang is a framework for running inference for large language models and multimodal models, meaning it processes inputs to produce model outputs such as text or other media. It is used to serve and accelerate open AI models and related workloads.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/sgl-project/sglangnpx agentmods add skills/sgl-project/sglang/sglang-diffusion-performanceWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/sgl-project/sglang/sglang-diffusion-performance)<a href="https://agentmods.dev/skills/sgl-project/sglang/sglang-diffusion-performance"><img src="https://agentmods.dev/badge/skills/sgl-project/sglang/sglang-diffusion-performance/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/sgl-project/sglang/sglang-diffusion-performance"><img src="https://agentmods.dev/badge/skills/sgl-project/sglang/sglang-diffusion-performance.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Snyk pass
- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00029 | $0.14245 |
| Opus 5.5 | $0.00012 | $0.05698 |
| Sonnet 5.5 | $0.00006 | $0.02849 |
| Haiku 4.5 | $0.00003 | $0.01425 |
Grade A, and why
sglang-diffusion-performance scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 15d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 566 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SGLang Diffusion Performance Tuning
Use this skill when the user wants the fastest command line, lower VRAM, or the right performance flags for a specific model and GPU setup.
Start from the model's cookbook page
Tune from the deployment we already recommend, not from scratch: read
docs/cookbook/diffusion/<Family>/ for the model first, take its command as
the baseline, then apply the options below. The cookbook page is what users
follow, so a recipe that silently disagrees with it on GPU count, parallelism
degrees, or offload flags is a recipe for a configuration nobody runs.
Families with a cookbook page: Cosmos, Ernie-Image, FLUX, Ideogram, JoyEcho, Krea, LTX, LingBot-Video, LingBot-World, LongLive, MOVA, MiniMax, Qwen-Image, SANA-Video, SANA-WM, Wan, Z-Image.
test/ configs are pinned CI setups for correctness and regression checks, not
deployment advice. Use them when you need an exactly reproducible run, and note
which cookbook flags you diverged from.
Before running any sglang generate command below inside the diffusion container:
- use
python/sglang/multimodal_gen/.claude/skills/sglang-diffusion-benchmark-profile/scripts/diffusion_skill_env.pyto derive the repo root, verify write access, and choose idle GPU(s) - export
HF_TOKENfirst when the selected model lives in a gated Hugging Face repo such asblack-forest-labs/FLUX.* - export
FLASHINFER_DISABLE_VERSION_CHECK=1 - when a run downloads weights, use a task-owned cache and delete that model's
cache after its eager/BCG/quality/profile group finishes; the benchmark
skill's
--quality-bcg-matrix --model-cache-root --cleanup-model-cachekeeps one cache for the group and writes a zero-residual cleanup ledger - hold one idle GPU set for the complete A/B matrix and verify no foreign process appears at run boundaries
cdto the repo root resolved fromsglang.__file__
Native Backend Gate
Performance numbers are useful only when the intended backend actually ran.
- Treat any log containing
Falling back to diffusers backend,Using diffusers backend, orLoaded diffusers pipelineas invalid for native SGLang performance tuning. - Use
--backend diffusersonly for an explicit diffusers baseline. For native recipes, leave the default backend or pin--backend sglang. - If a fallback happened, fix pipeline registration/model-path/config issues first, then rerun. Do not compare perf dumps collected from a fallback run.
- When the runtime auto-selects parallel settings because the user omitted them, keep the result as an auto-tuned baseline. For reproducible tuning, pin
--num-gpus,--ulysses-degree,--ring-degree, and--enable-cfg-parallelexplicitly.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 15d ago Changed 43938f6df6e7
- 20d ago Changed 78d106a19f82
- 25d ago Changed e78d3dd01115
- 29d ago Changed · +16 lines 740fca370b3b
- 1mo ago Changed · +88 lines 851c1d0cc0af
- 1mo ago Changed · +17 lines b009cc23179b
- 1mo ago Changed · +6 lines b14a4d8cca33
- 1mo ago First seen · 439 lines · 29 tokens per session scan A a843c76ffe45
sglang-diffusion-performance is a skill published in the GitHub repository sgl-project/sglang (37,623 stars, last pushed today), licensed Apache-2.0. It adds 29 tokens to every session and 14,245 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
llama-factory
Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2/3/4/5/6/8-bit QLoRA, multimodal support.
groq-inference
Ultra-fast LLM inference on custom LPU hardware. OpenAI-compatible API at api.groq.com. Lowest latency in the industry (500-1000+ tok/s). Supports chat completions, vision, audio (Whisper STT + TTS), tool calling, JSON mode, and streaming. Free tier available. Inference only — no training.
fireworks-ai-inference
Fast inference and fine-tuning platform with serverless and on-demand GPU deployments. OpenAI-compatible API for chat completions, embeddings, function calling, vision, and structured output. Supports SFT, DPO, and RL fine-tuning. SOC2 + HIPAA compliant.
huggingface-transformers
Use when running or fine-tuning Hugging Face Transformers models in Python — pipeline() inference for text classification and generation, loading weights with AutoModel/AutoTokenizer, devicemap and dtype memory management, tokenizer usage, Trainer fine-tuning, and BGE embeddings for Chinese RAG. Covers transformers…
mlx-serve
Hook an app, game or script up to the local mlx-serve server for LLM chat, embeddings, image, speech, music, sound effect, video and 3D generation, and Laya/Kev/Clef typed decisions, or make a whole film or music video on this Mac. Use when code should call mlx-serve or the user wants a video, song, pictures or game…
hqq-quantization
Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.