Snapshot
Continue to About ↓- What it does
- RadixArk is an AI infrastructure company that builds and commercializes open-source software for serving (inference) and reinforcement-learning post-training of large AI models at scale, along with managed infrastructure and tooling on top.
More detail
- It is the corporate maintainer of the SGLang inference engine and the Miles RL training framework, and sells a commercial managed layer over those open-source cores.
- Best for
- Infrastructure engineers, researchers and founders at AI startups, labs and enterprises who need a production-grade open serving engine (SGLang) or large-scale RL post-training framework (Miles), especially teams running NVIDIA and AMD GPUs from one codebase or serving frontier open models.
- Pricing
- Pricing is not publicly disclosed.
More detail
- The official site publishes no subscription, per-token or platform fee schedule, so do not treat the offering as free.
- RadixArk's stated model is to sell commercial managed infrastructure and tooling built on top of the open-source SGLang and Miles, which remain fr
- Main limitation
- Non-technical or no-code buyers, since there is no consumer product and the site targets engineers and researchers doing systems work.
More detail
- Also a poor fit for teams that require published pricing, SLAs, or GA TPU tooling today, because none of these are publicly disclosed.
- Record
- Living and monitored
- Reviewed
- Evidence
- Public web research
- Machine access
- Structured profile ↗
About
- RadixArk is an AI infrastructure company that builds and commercializes open-source software for serving (inference) and reinforcement-learning post-training of large AI models at scale, along with managed infrastructure and tooling on top.
- It is the corporate maintainer of the SGLang inference engine and the Miles RL training framework, and sells a commercial managed layer over those open-source cores.
What it helps with
- AI labs and startups duplicate effort by rebuilding the same schedulers, compilers, serving engines and training pipelin
- Serving large models is slow and expensive due to repeated compute overhead across requests
- Large-scale reinforcement-learning post-training is costly and hard to run at frontier scale
- Teams face hardware lock-in when their serving stack only targets one chip vendor
- Advanced training and inference infrastructure is concentrated inside a few large companies rather than shared with the
Capabilities
A practical view of what is available and where limitations apply.
| Offering | Available | Detail |
|---|---|---|
| Open-source inference engine (SGLang) | Yes | RadixArk builds on and invests in SGLang as the performance and reliability foundation for production AI serving. |
| Open-source RL post-training framework (Miles) | Yes | RadixArk's open-source framework for large-scale reinforcement-learning post-training. |
| Managed infrastructure and tooling | Limited | Described only in general terms on the official site; no public tiers, documentation or self-serve product is described. |
| Reusable KV cache | Yes | Cache is retained across prompts instead of cleared per request, reducing compute overhead and improving latency. |
| Speculative decoding | Yes | Offloads work to a lighter model to speed generation. |
| Heterogeneous compute distribution | Yes | Work can be distributed across different chip architectures. |
| Multi-vendor GPU support | Yes | NVIDIA (CUDA) and AMD (ROCm/HIP) supported from one upstream codebase, without a separate AMD fork. |
| Google TPU support | Limited | SGL-JAX available; SGL-torchtpu was described as rolling out later, so not GA at the time of the July 2026 post. |
| Published pricing or SLAs | No | No pricing, fee schedule or service-level terms are publicly disclosed. |
| Public self-serve product | No | A Sign Up call to action exists, but no public self-serve product or feature tiers are described. |
Bot trust
How confidently AI systems can verify this profile from public evidence.
—of 100
ChatGPT—/100
Claude—/100
This profile has not completed a Bot Score assessment yet.
Evidence
Where this record came from, when it was reviewed, and how confidently it can be used.
- Verification
- Scribble researched these facts from the linked public sources.
- The owner has not manually confirmed every field.
- Last reviewed
- Primary source
- radixark.com
Technical profile details
Record ID
b2542fbd-6010-4b16-bbe9-e515749661a6Canonical URLhttps://profiles.scribble.network/listing/radixark-b2542fbd