Revision History
Large Language Model · 20 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 a video.
Version 20 · Nov 1, 2026, 08:10 AM
- Correction: GLM-5.2 and GLM-5.3 sizes changed from an unsourced ~750B to Z.ai's 744B total and 40B active (model index), noting that the GLM-5.3 launch post calls the base model 743B; GLM-5.3's active parameters are 40B, not undisclosed; fixed the citations on those claims.
Version 19 · Oct 30, 2026, 09:46 AM
- Correction: the original GPT-4 had an 8,192-token context window (128,000 was GPT-4 Turbo), GPT-4.1's window is about 128 times GPT-4's rather than eight times, Gemini 3.5 Pro was never released, and Meta's 2026 frontier models are not open-weight. Added current API prices for 18 models, independent benchmark results for 14 models, an October 2026 frontier-model table, and 2026 updates for each major lab.
Version 18 · Oct 30, 2026, 06:45 AM
- Start the article with its lead paragraph: moved the first paragraph above the "What is a large language model?" heading. No text changed.
Version 17 · Oct 29, 2026, 09:24 AM
- Added a licensed causal language modeling illustration
Version 16 · Oct 2, 2026, 11:05 AM