An Unverified Tweet Is Moving AI Market Odds—Treat Anthropic's 'Next Week' Model as Rumour

The report is candid about what it is: a speculative post by a single social-media user (@cryptopunk7213) suggesting Anthropic could soon release a new, more affordable model that surpasses OpenAI's GPT-5.6 Sol in both intelligence and cost. There is no official confirmation from Anthropic. The piece notes Anthropic's current leading model is Claude Fable 5, available globally since 1 July following an easing of export controls, and that OpenAI's GPT-5.6 Sol has been leading recent benchmarks while priced competitively. A linked prediction market puts high odds on Anthropic holding a top-three spot by end of July.

We should be clear-eyed rather than breathless. A rumour is not a launch, and "surpasses GPT-5.6 Sol" is exactly the kind of framing we've learned to verify against hard benchmarks before repeating it. What's genuinely interesting is not the claim but the mechanism around it: a prediction market has turned an unverified tweet into a priced signal, with a business (Vera) monetising the "analysis" of that signal. That is a small case study in how AI hype now feeds financial products directly.

Our reading: on the fundamentals, the rumour is plausible precisely because it fits the pattern we already track—Anthropic and OpenAI trading the frontier lead while pushing prices down, which is good news for anyone who actually uses these tools. Cheaper, better models are the democratising engine of the whole story. But plausible is not confirmed. The thing worth watching is an official Anthropic announcement and independent benchmark results; until then, this is a market pricing a mood, not a model. When the numbers land, we'll judge them on the numbers.