DeepSeek has open-sourced every layer of an AI lab, from weights to inference engine
examachine · x · 2026-10-22
A recap of everything DeepSeek has open-sourced: MIT-licensed model weights, FlashMLA attention kernels hitting 580 TFLOPS on H800, the DeepEP MoE communication library, DeepGEMM FP8 matmul at 1,350+ TFLOPS, the 3FS distributed file system, the DeepInfer inference framework powering V4.1 Flash, and the deepseek-recipe deployment guide so self-hosted setups match their API. As the author notes, that covers model, kernels, communication, storage, and serving — everything but the training data.
More from Infra
- DeepSeek V4.1 Flash hits 1,521 tok/s on Fireworks Ultrafast, ships Tetris in 11 seconds — nicolechirps · 2026-10-22
- Citi projects Big Four hyperscaler capex to hit $1.62T annually by 2030, $5.35T cumulative — Beth_Kindig · 2026-10-22
- NVIDIA open-sources SimReady Central: 2,000+ sim-ready 3D assets on Hugging Face — lhoestq · 2026-10-22
- From devrel to CTO: juggling 12 cloud and AI providers and still running out of compute — NERDDISCO · 2026-10-22
- Coinbase's x402 was built to kill API keys; Stripe ships Machine Payments SDK for agent payments — kleffew94 · 2026-10-22
- AI software is only ~5% of enterprise IT budgets, says analyst Ben Bajarin — and compute is still constrained — BenBajarin · 2026-10-22