Will Google be the second-best Code Arena | WebDev AI lab at the end of December 2026?

The market resolves on which AI lab ranks second in the Code Arena | WebDev leaderboard at the end of December 2026. Google is currently priced at just 4%, implying a near-zero probability. Code Arena is a dynamic benchmark where positions can shift quickly with new model releases, and the horizon to end-2026 is very long. The low price reflects inertia and underestimation of long-term competition rather than any real weakness at Google.

Key drivers behind our estimate: Google has one of the strongest web-development teams (Gemini, AlphaCode, internal tooling), ships frequent updates, and has the resources to optimize specifically for benchmarks. Historically, Google has repeatedly held top positions in similar rankings, and the gap between first and second is often small. At 4%, the market assumes Google either will not compete or will lose to less-established rivals, which contradicts its track record and budgets.

For the outcome to win, Google must finish exactly second by December 2026, not first or third. That requires another lab (e.g., OpenAI or Anthropic) to take first while Google beats the rest. Given the broad field and long runway, we estimate this scenario at 30%, a substantial edge over the price.

Main risks: Google could dominate and take first place, nullifying the bet; or new entrants (xAI, Meta, startups) could push it out of the top two. Changes to Code Arena's methodology or discontinuation of the ranking are also possible. If Google fails to ship a competitive model by end-2026, the idea fails.