Revision History
Context caching · 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: Corrected OpenAI's pre-GPT-5.6 cache discount (model-dependent, 90 percent off on GPT-5, not a flat 50 percent), the origin of 24-hour retention (GPT-5.1, not GPT-5), the stale OpenAI TTL default, a wrong Gemini storage break-even calculation, a false claim that ai.google.dev omits storage pricing, an invented PromptHub savings figure, and a misdated, misdescribed XDA report on Claude Code's cache TTL; updated Gemini model rows to November 2026 (plus 26 minor fixes)
Version 7 · Nov 4, 2026, 08:53 PM
- Correction: fixed Claude Opus 4.7 cache pricing ($5 input, not $15), updated Anthropic minimum-token tiers (512 tokens for the 5.x models; Bedrock no longer differs), added OpenAI's GPT-5.6+ cache-write rates and explicit breakpoints, removed unsupported OpenAI '80% latency' quotes, corrected the Claude 3 Opus launch-support detail, and reconciled Google's May/June 2024 context-caching dates
Version 6 · Oct 29, 2026, 09:49 PM
- Improve internal links (2026-10-13)
Version 5 · Oct 13, 2026, 10:06 AM
- Added 6 contextual internal links
Version 4 · Jul 23, 2026, 03:45 PM