What is Fable 5?

Fable 5 is an AI model from Anthropic referenced in comparisons of performance, cost, and the impact of restricting public access to advanced models.

• Jun 2026 - Nathaniel Whittemore said GLM 5.2 beats Fable 5 on reasoning at 1/10 the cost and 300 tokens per second, and that Designer Arena reported GLM-52 beat Fable 5 at website design.
• Jun 2026 - Nathaniel Whittemore said stopping models like Fable 5 or Mythos 5 from being served to the public does nothing to slow down development and may speed it up by freeing up resources.
• Jul 2026 - Andrew Curran said many less capable models including Claude Opus 4.8, GPT 5.5, and Kimmy K2.7 could identify the same vulnerabilities as Fable 5 did in a report, and that Fable 5 frequently accepted pushback while sticking to its guns on other parts; in his limited experience, Fable 5 blew GPT55 and Opus 48 out of the water.
• Jul 2026 - Nathaniel Whittemore said the US government used an export control directive to demand that Anthropic suspend Fable 5 and Mythos 5 access for foreign nationals, and that Anthropic uses different guardrails on Opus than on Fable 5, expecting 85% fewer refusals.
• Jul 2026 - Nathaniel Whittemore said Anthropic had removed 80% of the system prompt for Opus 5 and Fable 5 and Claude Code, resulting in zero change to their coding benchmarks; he also said Opus 5 is nowhere near Fable in practical use, yet Opus beats Fable on many benchmarks, and that Fable scored 43.3% on Frontier Bench, about 10 points higher than Fable 5 and 9 points higher than GPT-5-6-Soul.
• Jul 2026 - Tariq said Fable is the first model where the quality of work is bottlenecked by his ability to clarify its unknowns; Matt Schumer said if you tell Fable to make something high quality, it stops at its own idea of good enough, which is usually lower.