Cacheon launches GLM-5.3 kernel arena, paying up to 33 TAO daily to beat sglang
JosephJacks_ · x · 2026-09-08
Inference-optimization competition platform Cacheon announced a GLM-5.3 arena where miners can submit kernel optimizations and earn protocol rewards.
How it works:
- Write Triton or CuteDSL GPU kernels targeting ops, fused blocks, or collectives in a fixed model
- sglang is the baseline: beat it at equal fidelity (KL fidelity gate, no benchmark regression) to earn a crown
- Rewards scale with speedup, up to 33 TAO per day, paid in SN14 tokens
Early results: RMSNorm kernel +37%, collective op +12%, attention block still open. Cacheon frames it as a path to production use and monetization of the subnet.
More from Infra
- Walking the AI rack optical stack: InP substrates and silicon photonics as the cleaner bet — demian_ai · 2026-09-08
- Running dual RX 7900 XTX on X570/X870 Taichi for local LLM inference: is x8/x8 enough? — espece-de-bon · 2026-09-08
- Hugging Face teases WebGPU inference engine with 5-10x speedups on Transformers.js — nicodotdev · 2026-09-08
- Meta to deep-dive recommendation inference systems at PyTorch Conference 2026 — PyTorch · 2026-09-08
- Memory crunch hits home: 4TB portable SSD prices stun as AI reprices the storage stack — demian_ai · 2026-09-08
- Sol-H3: MiniMax-H3 video generation faster than playback at 1.653s per 5s clip on 8x B300 — MiniMax_AI · 2026-09-08