AI Models Ranked — Why Every Lab Claims They're #1 | AI News

Claude Fable 5 leads the composite leaderboard at 100 out of 100. OpenAI claims GPT-5.6 Sol beats it — but nobody outside twenty approved partners can verify. Chinese models cost eighteen cents per million tokens against four dollars for the Western leaders. The benchmarks everyone cites are saturated, the pricing gap is wider than the performance gap, and the FTC is calling it AI washing.