🚨 GPT-5.6 Sol and Claude Fable 5 are now separated by just ONE point

• GPT-5.6 Sol xHigh — 1,631 • Claude Fable 5 — 1,630 • GLM-5.2 Max — 1,581 • Grok 4.5 — 1,557

The benchmark evaluates real frontend development and multi-step coding workflows rather than isolated coding questions.

Has OpenAI finally caught Claude for coding in your opinion?