Episode

ZAYA1-8B: Zyphra's Reasoning MoE Trained Entirely on AMD MI300X That Punches Far Above Its Weight Class - May 8, 2026

Podcast
DX Today | No-Hype Podcast & News About AI & DX
Published
May 8, 2026
Duration seconds
756
Processing state
not_requested
Canonical source
https://www.buzzsprout.com/2207817/episodes/19145129-zaya1-8b-zyphra-s-reasoning-moe-trained-entirely-on-amd-mi300x-that-punches-far-above-its-weight-class-may-8-2026.mp3
Audio
https://www.buzzsprout.com/2207817/episodes/19145129-zaya1-8b-zyphra-s-reasoning-moe-trained-entirely-on-amd-mi300x-that-punches-far-above-its-weight-class-may-8-2026.mp3
JSON
/v1/public/podcasts/dx-today-no-hype-podcast-news-about-ai-dx-6434212/episodes/zaya1-8b-zyphra-s-reasoning-moe-trained-entirely-on-amd-mi300x-that-punches-far-above-its-weight-class-may-8-2026
Markdown
/podcast/dx-today-no-hype-podcast-news-about-ai-dx-6434212/zaya1-8b-zyphra-s-reasoning-moe-trained-entirely-on-amd-mi300x-that-punches-far-above-its-weight-class-may-8-2026.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/dx-today-no-hype-podcast-news-about-ai-dx-6434212/episodes/zaya1-8b-zyphra-s-reasoning-moe-trained-entirely-on-amd-mi300x-that-punches-far-above-its-weight-class-may-8-2026/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/dx-today-no-hype-podcast-news-about-ai-dx-6434212/zaya1-8b-zyphra-s-reasoning-moe-trained-entirely-on-amd-mi300x-that-punches-far-above-its-weight-class-may-8-2026.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Send us Fan Mail ZAYA1-8B: Zyphra's Reasoning MoE Trained Entirely on AMD MI300X That Punches Far Above Its Weight Class - May 8, 2026 Zyphra dropped ZAYA1-8B this week, a sub-billion-active-parameter mixture of experts reasoning model pretrained end to end on AMD Instinct MI300X GPUs that matches DeepSeek R1 on competition mathematics and approaches Claude 4.5 Sonnet under their novel Markovian RSA test time compute. Chris and Laura unpack the architecture innovations (Compressed Convolutio...