Claude Fable 5.1, Opus 5, GPT-6 Astra and Gemini 3.8 Flash.

A busy few weeks for the model picker. Four new models are live in Morrison: Claude Fable 5.1, Claude Opus 5, GPT-6 Astra and Gemini 3.8 Flash. Same Skills, same integrations, same crawl on every one - switchable per message, as always.

Claude Fable 5.1

Anthropic’s flagship, and the successor to the Fable 5 we added in June. This is the model for the hardest agentic work: a full-site content audit that has to hold hundreds of pages and three data sources in its head, a decline diagnosis across noisy data, a long automation that chains a dozen tool calls before it writes a word. Where Fable 5 was already the one you reached for when judgment mattered, 5.1 is sharper on the long, multi-step runs.

Claude Opus 5

Near Fable-level reasoning at Opus pricing. Opus 5 replaces Opus 4.8 as the heavy-analysis option, and it is the sensible default when you want a flagship-class answer without spending flagship-class credits on every turn. Think programmatic-template audits, cannibalization calls, anything you would want a second opinion on.

GPT-6 Astra

OpenAI’s new flagship joins the GPT-5.6 family at the top of the OpenAI lineup, built for the hardest reasoning, research and coding. If your team already thinks in OpenAI models, this is the one to hand the difficult question to.

Gemini 3.8 Flash

The new Flash workhorse, and the default for new workflows. Stronger at agent loops than 3.7, still the light, fast, cheap choice for the quick pass over many pages - the first read before you hand the confusing cases to a flagship.

The lineup today

Claude Sonnet 5 remains the default for new chats. Above it sit Opus 5 and Fable 5.1. GPT-6 Astra leads the OpenAI side, with Sol, Terra and Luna below it. Gemini 3.8 Flash and Gemini Pro cover the Google side. Earlier models stay available for existing chats, marked as legacy in the picker, so nothing you built breaks.

Try it

All four are live in every workspace. Click the model name next to the send button and they are in the list, each with a one-line description and a cost tier. Run a fast pass on 3.8 Flash, hand the hard case to Fable 5.1, and see the difference in the same conversation.

Tell us what each one is earning its place on at hello@morrison.app.