Revision History
Cerebras Systems · 16 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.
- Update: added an official video.
Version 16 · Nov 1, 2026, 08:08 AM
- Correction: Cerebras Inference does not stream weights via MemoryX (weights stay resident in SRAM, split across systems); founders table titles updated (Sean Lie is CTO, Gary Lauterbach CTO emeritus, Michael James Chief Architect, Advanced Technologies); intro speed superlative now attributed with a dated Artificial Analysis measurement. Update: new subsection on how Cerebras serves multi-trillion-parameter models (Kimi K3 59-wafer worked example, pipeline concurrency, KV cache ceiling) and Oracle OCI plan.
Version 15 · Oct 30, 2026, 09:56 AM
- Improve internal links (2026-10-13)
Version 12 · Oct 13, 2026, 10:07 AM
- Correction: the Ultrafast section stopped at the August 13, 2026 Cerebras-powered GPT-5.6 Sol preview, letting a reader infer that Cerebras also powers the GPT-6 Astra Ultrafast tier that shipped on September 29, 2026. Added one sourced sentence and two sources.
Version 11 · Oct 1, 2026, 12:05 PM
- Added AMD Helios partnership, disaggregated inference strategy (prefill-decode, not attention-FFN), AWS Commercial Agreement and Bedrock timeline, WSE-3 Turbo and CS-4, Q1/Q2 2026 results, Europe and Finland data centres, IPO closing proceeds, and a September 2026 status paragraph
Version 10 · Sep 7, 2026, 03:23 AM