All LMArena props resolve WITH style control unless otherwise stated, resolves to the next Sonnet Model, I used 5 for clarity, but any # counts
Update 2026-02-03 (PST) (AI summary of creator comment): For image generation tools: Must be a dedicated diffusion model (like Midjourney/Stable Diffusion), not general tools like GIF creation skills.
Update 2026-02-03 (PST) (AI summary of creator comment): This market will resolve based on default settings with style control enabled on LMArena.
Update 2026-02-03 (PST) (AI summary of creator comment): Creator may resolve the market at 10 PM ET instead of waiting until the official close date if Claude Sonnet 5 is not released within the next hour or two from the time of the comment.
Update 2026-02-03 (PST) (AI summary of creator comment): Creator will wait until the official close date to resolve the market, rather than resolving early at 10 PM ET on February 3rd.
Update 2026-02-04 (PST) (AI summary of creator comment): The creator will defer to @MarryBobinson's preference on how to resolve the "Releases before 3pm Feb 4th" answer (despite the typo "reseases").
1,000People are also trading
CLARIFICATION: the times are in PST for my markets, as that's where anthropic is
@MarryBobinson how can it reseases? I think that release and reseases are two very different things.
Therefore, it should resolve no.
@Velaris the convention on manifold is to resolve according to the obvious intent when intent and literal writing differ
@JeromeHPowell Hopefully, that seems fair. I just don't want to get caught in the crossfire of resolution wars
@Bayesian Yeah, great q, I'm not sure tbh, it would double opus which seems like a lot
@JeromeHPowell imo more relevant data is: sonnet 4.5 released with 128k iirc, but then around a month later they released a 1M context version
@Bayesian What if it is 200k or whatever in web chat / Claude Code, but longer available only via API?
@EvanDaniel whichever is longer is what the model’s capability is at. Client usually is lower than api bc artificial cost saving measure
@EvanDaniel Resolves YES if a >400k option is available at all, doesn’t need to be in Claude Code or web chat (though I expect the main selling point of long context will be reduced compaction frequency in Claude Code)