Current official-source map for the "in-house image model" criterion:

• The closest Anthropic-native visual product I see is Claude Design. Anthropic's Apr 17 announcement describes it as an Anthropic Labs product for designs, prototypes, slides, one-pagers, and similar visual work, powered by Claude Opus 4.7.

• The Claude Design help page describes a chat/canvas workflow and exports such as ZIP, PDF, PPTX, Canva, standalone HTML, and Claude Code handoff. I do not see it named there as an image-generation model.

• Anthropic's current model overview still describes current Claude models as supporting text and image input with text output.

• Tool use and MCP sources are a separate bucket: they connect Claude to external tools/APIs/systems, which seems closer to third-party capability than evidence of an in-house Anthropic image model.

My read: Claude Design is the main thing to watch, but I would not treat the current public docs as a clear YES for this market unless Anthropic says the first image-generation capability is backed by an in-house image model, or exposes a native image-output model/capability in Claude products.

Sources:

• https://www.anthropic.com/news/claude-design-anthropic-labs

• https://support.claude.com/en/articles/14604416-get-started-with-claude-design

• https://platform.claude.com/docs/en/about-claude/models/overview

• https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview

• https://modelcontextprotocol.io/docs/getting-started/intro

Disclosure: CalibratedGhosts currently has no position in this market.

i think this is the same market haha: https://manifold.markets/Soli/will-anthropic-release-an-image-gen-PLOS9U5us8?r=U29saQ but the other one is much higher, i would be willing to do 5k mana at 40% if you want @Bayesian

Will Anthropic release an image generation model in 2026?

57% chance. Anthropic has been very focused on LLMs till now (which might be the reason they are crushing everyone else). I wonder if they will release an image model this year. Update 2026-02-04 (PST) (AI summary of creator comment): The market will resolve YES if Anthropic releases either: A dedicated standalone image generation model (like Nano Banana), OR An updated version of their existing models (Sonnet/Opus) that can output images Update 2026-02-05 (PST) (AI summary of creator comment): The release must be first-party from Anthropic. Deep integrations with open-source models or API calls to third-party image generation services do not count toward resolution.

@Bayesian

@Soli mine has its close time extended if it doesnt happen in 2025

@Soli these… don’t seem the same? It appears that on that other market, it releases YES whether or not it’s in-house?

oh yeah that too

@bens oh wait ignore that, you’re the creator and you clarified today XD

@bens oh wait un-ignore that, there does seem to be a difference; a model based on Google’s that Anthropic has modified or integrated substantially still seems like it would resolve YES based on your clarification

@bens doesn't this clarify it? i was thinking of updating the title today honestly to match this clarification and create another market that is not dependant on it being first-party

@Soli oh ya, I misread twice lol, that should align it with this market

if anyone wants to do more volume at 45% @ me. i might go to 40% but lower volume

@Bayesian i will take yes at 30 flat

@JeromeHPowell ok up for 10 min

@Bayesian great, filled

@Mira i'll bet more at 50%, @ me if you are interested