Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
Simon Willison's same-day roundup: Opus 5.5 ships, then GPT-6 Sol and Luna about an hour later, framed as a new price war.
AI Storyline
7 items · 6 sources · 6 days
Operational story trace
Opus 5.5
Follow in this browser to see new updates on your Live feed.
Current statePrice war · multi-lab cuts
Latest change
Claude Sonnet 5.5 reached Amazon Bedrock and Claude Platform on AWS on Sep 28, pitched at a lower cost per task and faster speed than earlier Sonnet models.
Earlier contextThe story so far
Anthropic cut Fable/Mythos 5.1 cache prices 75% in early September after Zhipu's GLM-5.3-Flash undercut. On Sep 22 Opus 5.5 and OpenAI's GPT-6 Sol and Luna shipped within an hour of each other, and 40-50% price cuts spread across labs.
editor-curated · source-linked
State over time
● cache price cut · Sep 2Sonnet 5.5 on AWS · Sep 28 → now ●
- cache price cut · Sep 2
- dueling launches · Sep 22
- price cuts spread · Sep 23
- cost guidance · Sep 24-25
- Sonnet 5.5 on AWS · Sep 28 → now
LAUNCH · Sep 22Anthropic ships Claude Opus 5.5; OpenAI answers within the hour with GPT-6 Sol and LunaSame-day flagship releases from both labs, framed as a new price war rather than a one-off launch.1 source · show source ▾
FALLOUT · Sep 23Everybody cuts prices 40-50% in Opus 5.5's wake1 source · show source ▾
[AINews] Claude Opus 5.5, the new default model for AINews — and everybody cuts prices 40-50%
AINews confirms the launch triggered board-wide 40-50% price cuts, not just a two-lab skirmish.
ROLLOUT · Sep 22–25Opus 5.5 lands on AWS Bedrock, then gets its own cost-and-context guidance3 sources · show sources ▾
Claude Opus 5.5 is now available on AWS
Opus 5.5 reaches AWS Bedrock and Claude Platform on AWS, broadening where builders can deploy it.
Coding sessions are longer and use more context. Claude Opus 5.5 is built with that in mind. | Claude by Anthropic
Anthropic's own post: coding sessions are longer and use more context now, and Opus 5.5's design and pricing account for that.
What a task costs on Opus 5.5 / claude.dev Blog
Anthropic's own guide: what a Claude Code task actually costs on Opus 5.5, and which settings move the bill.
SONNET · Sep 28Sonnet 5.5 lands on AWS as the lower-cost-per-task option1 source · show source ▾
Introducing Claude Sonnet 5.5 on AWS
AWS announces Sonnet 5.5 on Bedrock and Claude Platform on AWS, described as lower cost per task at faster speed; also covered by Claude's dev blog and Simon Willison.
NOW · Sep 29Early hands-on: Opus 5.5 handles explainer videos1 source · show source ▾
[AINews] Opus 5.5 is good at explainer videos
AINews' Sep 29 issue: early hands-on report that Opus 5.5 is good at explainer videos; no pricing or availability change.
Day 1 Tuesday, Sep 22, 2026
Claude Opus 5.5 is now available on AWS
Opus 5.5 reaches AWS Bedrock and Claude Platform on AWS, broadening where builders can deploy it.
Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
Simon Willison's same-day roundup: Opus 5.5 ships, then GPT-6 Sol and Luna about an hour later, framed as a new price war.
Day 2 Wednesday, Sep 23, 2026
[AINews] Claude Opus 5.5, the new default model for AINews — and everybody cuts prices 40-50%
AINews confirms the launch triggered board-wide 40-50% price cuts, not just a two-lab skirmish.
Day 3 Thursday, Sep 24, 2026
Coding sessions are longer and use more context. Claude Opus 5.5 is built with that in mind. | Claude by Anthropic
Anthropic's own post: coding sessions are longer and use more context now, and Opus 5.5's design and pricing account for that.
Day 4 Friday, Sep 25, 2026
What a task costs on Opus 5.5 / claude.dev Blog
Anthropic's own guide: what a Claude Code task actually costs on Opus 5.5, and which settings move the bill.
Day 5 Monday, Sep 28, 2026
Introducing Claude Sonnet 5.5 on AWS
AWS announces Sonnet 5.5 on Bedrock and Claude Platform on AWS, described as lower cost per task at faster speed; also covered by Claude's dev blog and Simon Willison.
Day 6 Tuesday, Sep 29, 2026
[AINews] Opus 5.5 is good at explainer videos
AINews' Sep 29 issue: early hands-on report that Opus 5.5 is good at explainer videos; no pricing or availability change.
What to watch — open questions
- Do Claude Opus 5.5's and GPT-6 Sol/Luna's posted prices actually beat GLM-5.3-Flash's domestic-chip pricing, or just each other?
- Which labs specifically matched the 40-50% cut, and does actual per-task cost (per AWS's guidance) net out cheaper despite Opus 5.5's higher list price?
- Does Sonnet 5.5's cost per task on Claude Code undercut Opus 5.5 by enough to change default model choice?
- Do further hands-on reports show Opus 5.5's per-task cost advantage holds outside coding workloads?
How this thread was built
editor wrote the arc · 5 beatswatcher 1 status change
Storylines are threaded mechanically from the feed: stories that share a distinctive anchor across multiple days and sources. Each item links to its original source. The evidence trace, current state, and open questions are written by the editor routine and refreshed whenever a new beat lands.