Revision History
Zyphra · 7 revisions
Sizes are character counts of the article source. The signed number is the change from the previous revision.
Recent edit summaries
Detailed summaries recorded by editors. Generic maintenance summaries are omitted here; every recorded revision remains below. Dates describe the edit, not necessarily the event it covers.
- Correction: fixed ZAYA1-8B benchmark attributions (the 88.3 HMMT'25 figure is GPT-5-High, not Claude 4.5 Sonnet; the 5.5M-token budget applies to APEX-shortlist), cluster size (1,024 GPUs, not nodes), Zamba v1 tokenizer, ZAYA1-VL comparator, unsupported Ultra Ethernet/memory claims, unverified founding-year and post-money details, and an unsupported claim of a London office (the company says it hires in London). Update: added ZAYA1-base, ZAYA1-74B-Preview, diffusion preview, Zamba2-VL, ZONOS2, ZUNA EEG models, Zyphra Cloud on MI355X, the reported $500M Series B talks (Forbes, May 2026), and 2026 research.
Version 6 · Oct 27, 2026, 07:24 AM
- Improve internal links (2026-10-13)
Version 5 · Oct 13, 2026, 10:03 AM
- Added 6 contextual internal links
Version 4 · Jul 23, 2026, 03:30 PM