---
title: 'In the News: October 9, 2026 (Extra 3)'
description: 'Sonnet 5.5 cache reads now cost $0.10 per million tokens, down from $0.20 at launch, and Codex adds an Ultrafast mode for GPT-6.1 Sol.'
canonical_url: 'https://darkfactory.dev/news/2026-10-09-extra-3'
markdown_url: 'https://darkfactory.dev/news/2026-10-09-extra-3.md'
collection: news
date_published: '2026-10-09T17:45:00-04:00'
date_modified: '2026-10-09T17:45:00-04:00'
---

# In the News: October 9, 2026 (Extra 3)


Anthropic has cut the price of Sonnet 5.5 cache reads in half since launch, and OpenAI has put a faster generation mode for GPT-6.1 Sol in Codex behind its top plans. Both change what an agent loop costs or how quickly it runs. Neither source reports a measured effect on a real workload.

## 1. Sonnet 5.5 cache reads are now $0.10 per million tokens, half the launch price

**[Claude Code changelog, version 2.1.296](https://code.claude.com/docs/en/changelog)** · Anthropic · October 9, 2026

The 2.1.296 entry says Claude Code's `/cost`, status line, `--max-budget-usd` and SDK cost figures now price Sonnet 5.5 cache reads at $0.10 per million tokens, and gives the old figure as $0.20. The 2.1.284 entry from September 28 listed the launch price as $2 per million input tokens, $10 output and $0.20 for cache reads. Anthropic's [pricing page](https://platform.claude.com/docs/en/about-claude/pricing) now shows $0.10 for cache hits and refreshes, with input and output unchanged. Neither page says when the new rate took effect or calls it a price cut, so the date of the change is unknown.

**Why it matters:** If your harness keeps a long context warm across many steps, the cache-read rate is the price-list line that governs that spend. It halved for Sonnet 5.5, so re-measure the per-run cost of your loops before you move work to a different model on price alone.

## 2. Codex gets an Ultrafast mode for GPT-6.1 Sol, limited to the $500 Pro plan and eligible business plans

**[ChatGPT and Codex changelog, October 8, 2026](https://learn.chatgpt.com/docs/changelog)** · OpenAI · October 8, 2026

The entry says Ultrafast mode speeds up token generation with GPT-6.1 Sol in Codex and ChatGPT Work. It is available on the Pro plan at $500 and on eligible Enterprise and Edu plans. Enterprise access is off by default, and workspace owners must turn it on. The entry says GPT-6.1 Sol Ultrafast supports inference residency in the United States and Europe (EEA and Switzerland). It gives no speedup figure and points to separate pages for usage and credit rates, which we did not read.

**Why it matters:** Generation speed is now a plan-gated option for this agent. Without a published speedup or credit rate, the only way to know whether it pays for itself is to time it on your own loop.