The weekly strip · 23 issues · page 2 of 2
Model watch
What shipped across the labs, with dates, prices and limits from the release notes.
- Issue 14
10 October 2026 · Model watch · 6 min read
The July 2026 OpenAI agent incident: what is reported, and what to change
What sources say about OpenAI's July 2026 agent incident, the UK AISI Astra tests and the GPT-6.1 shelving, plus a default-deny tool policy you can copy.
- AI safety
- agents
- OpenAI
- sandboxing
- tool permissions
- Issue 13
10 October 2026 · Model watch · 6 min read
GPT-6's Intelligent UI: what widget answers mean for product teams
ChatGPT answers now arrive as buttons, charts and forms. What OpenAI has said, what it has not, and how to design chat features and Apps SDK widgets around it.
- GPT-6
- Intelligent UI
- generative UI
- product design
- MCP Apps
- Issue 12
10 October 2026 · Model watch · 7 min read
GPT-6 Astra, Sol, Luna and Instant: what is published and what is not
GPT-6 has a published API price list, but Instant is not on it. What OpenAI's pricing page, press and Wikipedia agree on, where they conflict, and how to plan around it.
- GPT-6
- OpenAI
- model selection
- LLM pricing
- GPT-6 Astra
- Issue 11
10 October 2026 · Model watch · 6 min read
Gemini 2.5 is limited, not deprecated: what that means for your project
Since 18 September, Gemini 2.5 access is limited to prior users. The text models have no shutdown date; the audio and image ones do. Dates, prices and a migration check.
- Gemini 2.5
- deprecation
- migration
- Gemini API
- implicit caching
- Issue 10
10 October 2026 · Model watch · 6 min read
Nano Banana 2.1 in the Gemini API: 4K pricing and where it fits
gemini-nano-banana-2.1 went GA on 6 October with 1K, 2K and 4K output and panoramic ratios. Per-image prices, what it replaces, and a Python call you can adapt.
- Nano Banana 2.1
- image generation
- Gemini API
- pricing
- Open Graph images
- Issue 08
10 October 2026 · Model watch · 6 min read
Gemini 3.8 Live extended thinking: what it buys a voice agent
Gemini 3.8 Live and its extended-thinking variant went GA on 15 September 2026. What the extra reasoning does, what it costs the client, and a latency budget.
- Gemini 3.8 Live
- voice agents
- Live API
- extended thinking
- latency
- Issue 07
10 October 2026 · Model watch · 6 min read
Gemini 3.8 Flash: 3.7 now routes to it, and what to test
Since 8 October 2026, requests to gemini-3.7-flash are routed to gemini-3.8-flash at the same price, until rates double on 1 January 2027. What to test before it bites.
- Gemini 3.8 Flash
- Gemini API
- migration
- LLM pricing
- model routing
- Issue 06
10 October 2026 · Model watch · 6 min read
Claude Sonnet 4.5 retires on 30 November 2026: a migration checklist
claude-sonnet-4-5-20250929 retires on 30 November 2026. The settings that now return a 400 on Sonnet 5.5, the cost delta, and how to test before the date.
- Claude Sonnet 4.5
- Claude Sonnet 5.5
- deprecation
- migration
- LLM pricing
- Issue 05
10 October 2026 · Model watch · 6 min read
Claude Opus 5.5 breaking changes: thinking can't be disabled
Opus 5.5 costs less per token than Opus 5 but returns a 400 for thinking disabled and forced tool_choice. The exact errors, the cost trap and a migration checklist.
- Claude Opus 5.5
- migration
- API
- adaptive thinking
- LLM pricing
- Issue 04
10 October 2026 · Model watch · 7 min read
Claude Haiku 5.5 for classification: what changed and what it costs
Haiku 5.5 launched 7 October 2026 at $0.10 input and $0.50 output per million tokens. What changed from 4.5, what returns a 400, and a classification bill priced out.
- Claude Haiku 5.5
- classification
- LLM pricing
- structured outputs
- prompt caching
- Issue 01
10 October 2026 · Model watch · 8 min read
Which AI model should you build on in October 2026?
Claude 5.5, GPT-6, Gemini 3.8 and TypeSafe's Jev all shipped in five weeks. Dates, prices and limits from primary sources, and what I would actually pick for a product.
- Claude 5.5
- GPT-6
- Gemini 3.8
- TypeSafe Jev
- model selection