RadixArk (@radixark) / X

RadixArk

113 posts

RadixArk

@radixark

SHIP AI FOR ALL.

Joined November 2025

Pinned
RadixArk
@radixark
10h
Very excited to partner with Google on TPUs! We're committed to delivering fast, native TPU inference for major open models, and support for advanced features. More news to come 🚀
Google for Developers
@googledevs
12h
Big news: @Google and @radixark are partnering to bring @sgl_project to Google Cloud TPUs! ✅ Run SGLang on TPU today via SGL-JAX ✅ Coming soon: SGL-torchtpu for a PyTorch-native experience ✅ No more migration tax—just ultimate flexibility for devs Learn more 🛠️:
RadixArk
@radixark
14h
RadixArk and Google Cloud are joining forces with the SGLang community to make TPU a drop-in, cost-efficient path to frontier inference. SGL-JAX already serves the major open model families on the latest TPU generations: Gemma, Qwen, DeepSeek, GLM, Kimi, Ling, MiniMax, MiMo,
RadixArk
@radixark
Jul 29
Excited to see Miles power the full post-training loop of Instella-MoE! Miles handles distributed rollouts, GRPO, and Ray-based multi-node orchestration, all running natively on AMD Instinct GPUs with ROCm. IFEval rises from 77.1 to 83.7 via IF-specialized RL + on-policy
pkms 🍕
@PrakamyaMishra
Jul 27
🚀 We are excited to introduce Instella-MoE✨, AMD's first fully open Mixture-of-Experts (MoE) language model! Instella-MoE has 16B total parameters with only 2.8B active parameters per token. Trained from scratch on AMD Instinct™ MI300X and MI325X GPUs using AMD-Primus and
RadixArk
@radixark
Jul 27
We trained a DSpark speculator draft model for Kimi K3 with SpecForge that takes batch-1 decode from ~113 to ~423 tok/s. Benchmark highlights: - +68% throughput at bs=256 on chat vs verify-all, lossless - 5.67 accept length on GSM8K, 5.36 on HumanEval Blog, serving command, and
00:52
LMSYS Org
@lmsysorg
Jul 27
SGLang day-0 speed on Kimi K3: 423 tok/s (measured on gsm8k), plus RL support ready in Miles @radixark! How the largest open-source model runs this fast: we natively implemented and deeply optimized K3’s new architecture with fused KDA decode kernels, DP attention, DSpark, PD
RadixArk
@radixark
Jul 27
RadixArk has proudly co-signed this letter. Our mission is to make frontier AI infrastructure open and accessible to everyone. We power the open-weight ecosystem with SGLang for serving and Miles for training. Excited to build the future of AI together with the community!
Jensen Huang
@JensenHuang
Jul 24
For my first post, I’m sharing a letter @nvidia signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
See @radixark's full profile
Sign up
Log in

See @radixark's full profile

See @radixark's full profile