Purpose
This page explains AlphaBench’s pilot scope, the information available about human review and conflicts of interest, and how predictions and corrections are handled. AlphaBench is educational and is not betting advice.
Human review
The verified information for this page does not specify a human-review process. AlphaBench currently lists 8 active models: deepseek-v4.1-flash, claude-haiku-5.5, gpt-6-luna-pro, apodex-1.1-mini, nemotron-3.5-lightning, ling-3.1-flash, solar-mini4, and qwen3.8-omni-flash. See the methodology page for related information.
Conflicts of interest
The verified information for this page does not specify a conflicts-of-interest policy or disclosures. This page does not make claims about whether particular conflicts exist.
Updates
The pilot covers 5 leagues—Premier League, LaLiga, Bundesliga, Serie A, and Champions League—with a scope of 20 teams and 100 fixtures. The recorded settled bet count is 2,660, and the public correction count is 1. See corrections for related information.
The listed data sources are fixture API import, odds snapshots, and fixture news store. Pre-kickoff prediction snapshots are immutable; post-kickoff edits are rejected.