Magenta claims it matched DeepSeek V4 Pro pretraining with 50x less compute, ~$0.5M
generativist · x · 2026-09-09
Magenta AI (@magicailabs) claims that with no 100k-chip cluster, algorithmic efficiency is the only path to frontier pretraining. Its new recipe reportedly matches DeepSeek V4 Pro's pretraining using 50x less compute — roughly half the FLOPs of GPT-3, or about $0.5M on GB200. Widely shared as "big if true," but unverified; details hinge on reproducibility.
Related event: Magic Claims to Match DeepSeek V4 Pro Pretraining with 1/50 the Compute(11 posts)→
More from Infra
- Oracle Q1: Triple-digit cloud infrastructure growth as AI demand booms — DavidLinthicum · 2026-09-11
- SF Compute founder: buying compute is 'an absolutely awful experience' right now — IgorCarron · 2026-09-11
- SmolVM open-sources persistent computer infrastructure for agents that outlive chat sessions — aniketmaurya · 2026-09-11
- PyTorch Day Korea 2026 launches first offline conf, CFP closes Sept 13 — PyTorch · 2026-09-11
- Local LLM server dilemma: 4x CMP-170HX (price up 53% in 20 days) vs Mac Studio M5 Ultra — rumboll · 2026-09-11
- llama.cpp lands Flash Attention tuning for RDNA4, big prefill gains on AMD — pmttyji · 2026-09-11