Writing benchmark: Astra ranks 33rd as Fable 5.1 takes first

Louis-François Bouchard 🎥🤖 @Whats_AIOriginally posted 2026.09.12

Open original on X

GPT-6 Astra vs. Fable 5.1, according to our internal writing benchmark. Best config of both. Fable wins for storytelling and voice. It's the best one of the benchmark. Fable 5.1 (max effort) by @AnthropicAI is #1 of 148 configs. GPT-6 Astra (max) by @OpenAI lands at #33. Fable wins 9 of our 10 tasks. But Astra follows the instructions just as well, and does it at $0.85 per script vs $3.15. 💸 More insights 👇