README: Opus 4.7 advisory, updated image cost for 4.7 tokenizer + high-res

- Top-level advisory: 4.7 burns Q5h at 2.4x rate of 4.6 (35% tokenizer +
  ~105% adaptive thinking). Links to Discussion #25 and workaround.
- Image stripping section: updated token costs for 4.7 (62.5K → 85K+ per
  500KB image due to tokenizer inflation + high-res support).
- Strongly recommended image stripping on 4.7.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: cnighswonger

README.md

@@ -4,7 +4,9 @@
 
 English | [中文](./README.zh.md) | [한국어](./README.ko.md) | [Português](./docs/guia-pt-br.md)
 
-Fixes prompt cache regressions in [Claude Code](https://github.com/anthropics/claude-code) that cause **up to 20x cost increase** on resumed sessions, plus monitoring for silent context degradation. Confirmed through v2.1.111. Opus 4.7 model supported.
+Fixes prompt cache regressions in [Claude Code](https://github.com/anthropics/claude-code) that cause **up to 20x cost increase** on resumed sessions, plus monitoring for silent context degradation. Confirmed through v2.1.111. Opus 4.7 compatible.
+
+> **Opus 4.7 advisory:** Our metered data shows 4.7 burns Q5h quota at **~2.4x the rate of 4.6** for equivalent visible token counts. Two factors: a new tokenizer (up to 35% more tokens, [documented](https://platform.claude.com/docs/en/about-claude/models/whats-new-claude-4-7)) and adaptive thinking overhead (~105%, not documented in usage response). Workaround: `CLAUDE_CODE_DISABLE_ADAPTIVE_THINKING=1` (may reduce quality). Image stripping (`CACHE_FIX_IMAGE_KEEP_LAST`) is even more important on 4.7 due to high-res image support increasing image token counts. See [Discussion #25](https://github.com/cnighswonger/claude-code-cache-fix/discussions/25) for full analysis.
 
 ## Security model
 
@@ -308,7 +310,7 @@ When the server downgrades your TTL to 5m (Layer 2 — quota-aware downgrade at
 
 ## Image stripping
 
-Images read via the Read tool are encoded as base64 and stored in `tool_result` blocks in conversation history. They ride along on **every subsequent API call** until compaction. A single 500KB image costs ~62,500 tokens per turn in carry-forward.
+Images read via the Read tool are encoded as base64 and stored in `tool_result` blocks in conversation history. They ride along on **every subsequent API call** until compaction. A single 500KB image costs ~62,500 tokens per turn on Opus 4.6, and potentially **~85,000+ tokens on Opus 4.7** due to the new tokenizer (up to 35% inflation) and high-res image support (2576px max, up from 1568px). Image stripping is strongly recommended on 4.7.
 
 Enable image stripping to remove old images from tool results: