Claude update appears to trigger compaction much earlier; context cap halves after tool use
Spiritual-Bathroom54 · reddit · 2026-10-16
A user's detailed testing suggests the new Claude update (likely tied to the Sep 16 Chat/Cowork merge, unconfirmed by support) compacts long chats far earlier. Using the unofficial Claude Usage Tracker extension, they found context max drops after artifacts/code execution: Sonnet 5/5.5 and Haiku 5.5 fell from 1M to 500K, Sonnet 4.6 and Opus 4.6 from 500K to 200K, while Opus 5.5 was unaffected. Post-update chats grew 6k tokens/turn vs 1.2k pre-update; one 264-turn chat consumed 36% of the weekly limit. Compaction also broke verbatim retention of 330KB reference files. The post includes a reproducible self-check method and asks others to compare.
More from Models
- OpenAI's Decisions API hits GA: near real-time routing up to 10x faster than GPT-6 Luna — OpenAIDevs · 2026-10-23
- DeepLearning.AI weekly: GPT-6.1 hits 97.9% byte parity on COBOL migration — DeepLearningAI · 2026-10-23
- Marker's lite mode scores 75.1 vs 42.3 for Azure Read on OmniParseBench — VikParuchuri · 2026-10-23
- Datalab launches lite OCR at $0.50 per 1,000 pages, 20-30x cheaper than Azure, Google and AWS layout APIs — VikParuchuri · 2026-10-23
- LiveNerf's 29-day data finds no nerfing of Claude Opus 5.5 — TheOnlyVibemaster · 2026-10-23
- PHP Foundation reveals Anthropic's Claude cyber model scanned PHP, validated with OpenAI's Daybreak Red — saibharadwaj · 2026-10-23