On Thursday, Alphabet (GOOGL.US, GOOG.US) overtook long-time favorite Anthropic in prediction market implied probabilities to claim the top spot, as traders bet on which tech giant will hold the strongest AI model by the end of this year.

According to Polymarket data, Alphabet currently leads with a 43% implied probability, with the market viewing it as most likely to release the best AI model before December 31, narrowly edging out Anthropic's 42%. OpenAI and Meta (META.US) trail significantly, with implied probabilities of 6.5% and 2.9%, respectively.

On the Kalshi platform, nearly identical shifts are unfolding: Alphabet's Gemini saw its implied probability rise to 41.9%, surpassing Anthropic's Claude at 41.3% and OpenAI's ChatGPT at 8.4%.

For most of this year, Anthropic held the leading position on both platforms, with implied probabilities frequently reaching 70% to 80%. But in early October, Anthropic's implied probability fell sharply, opening the door for Alphabet to climb from single digits and seize the lead as the year-end deadline approaches.

Earlier on Thursday, Alphabet announced Gemini Agent at its Gemini at Work 2026 event. This is a single general-purpose agent capable of answering questions, handling knowledge work, creating images and media content, and writing and running code. Shortly before that, Meta released Muse, and OpenAI released Dots.

Meanwhile, investors have been lowering expectations for this autumn's IPO wave. Due to market volatility and broader uncertainty, several companies have postponed their listing plans. As one of the most anticipated IPOs, Anthropic now aims to launch its offering before Thanksgiving, having previously considered going public earlier in the fall. Its main rival, OpenAI, is not expected to list until at least 2027.