Side-by-side comparison of 2-5 AI models. Returns pricing, benchmarks normalized to union-of-keys with explicit nulls, live provider status, recent news mentions, and per-benchmark rankings in one ready-to-rank payload.