Google DeepMind — Description: AlphaFold 2 achieved competitive-with-experimental-methods accuracy at CASP14 (Nov–Dec 2020), widely characterized as solving the protein folding problem for structural biology purposes

NoteThe source confirms the CASP14 competition timing (Nov-Dec 2020) and general scale (~100 groups, 90 targets). However, the source does not mention AlphaFold 2 by name, does not provide the median GDT score of 92.4, does not attribute leadership to John Jumper (it lists John Moult as CASP chair), and does not characterize any method as 'solving the protein folding problem.' The source is the official CASP14 website but appears to be a general informational page that does not contain the specific performance metrics or characterizations claimed. The claim's specific data points (GDT ~92.4, John Jumper leadership, 'solving' characterization) are unverifiable from this source, though the competition timing and general structure are confirmed.