Claude Fable 5 (Anthropic)

On June 30, 2026, Anthropic announced the redeployment of Claude Fable 5 and Claude Mythos 5 after US export controls imposed on June 12, 2026 were lifted. Fable 5 became available globally on July 1, 2026 via the Claude Platform, Claude.ai, Claude Code, and Claude Cowork, with Pro, Max, Team, and select Enterprise pla The export-control incident suspended access for all users because nationality could not be reliably verified in real time. Anthropic frames the resolution as a basis for standardizing how the industry triages jailbreak findings, launches highly capable models, and communicates risk to government and industry partners.

A June 10, 2026 third-party developer guide confirms Claude Fable 5 shipped June 9, 2026 as the first public model from the same family as Mythos, sitting one tier above Opus 4.8. Fable 5 is shipped with three classifier categories bolted on that, when tripped, cause the model to decline; those refusals fall back to Op The guide covers model IDs, pricing, API calls, and Claude Code integration, with figures checked against Anthropic's June 2026 documentation and run live against the claude-fable-5 endpoint. It frames the Fable/Mythos split as Anthropic's way of bringing Mythos-level capability to the public without handing everyone a

Updated September 5, 2026, this deep-dive covers the June 9, 2026 release of Claude Fable 5, framed by Anthropic as the most powerful model it offers generally, built for long-running, complex work. Reported capabilities include 12-hour autonomous runs and a multi-million-line code migration completed in a single day, The article logs the June 12, 2026 US government directive that suspended Fable 5 and Mythos 5 for all users and the July 1, 2026 redeployment about 19 days later. Specs and figures are quoted from Anthropic's official announcement and various reports as of June 2026, with availability and pricing noted as subject to c

Published June 9, 2026, this review aggregates Fable 5's core specifications: model ID claude-fable-5 on the Claude API, Mythos-class tier one step above Opus, $10 per million input tokens and $50 per million output tokens, 1M input tokens / 128K output tokens context window, text plus vision input with text output, an Reported benchmark numbers include 95.0% on SWE-bench Verified, 80.0% on SWE-bench Pro, a leading 1932 Elo on GDPval-AA, and a top rank on FrontierCode, with pricing framed as double Opus 4.8 but less than half what Mythos Preview cost. The review stresses that the safeguard layer matters more than any single benchmark

Published June 9, 2026, this governance-oriented analysis frames the Fable 5 / Mythos 5 split as a capability-governance architecture rather than a simple model release. Fable 5 is the public configuration for developers, while Mythos 5 is a reduced-safeguard variant for vetted government and research partners via Proj The article clarifies that all benchmark figures are vendor-reported, noting SWE-bench Pro and SWE-bench Verified are different benchmarks whose results are not directly comparable. Reported figures include SWE-bench Pro at 80.3%, Hex Core Analytics at 90%, and Cognition FrontierCode Diamond at 29.3%. It argues Project

Published June 9, 2026, this technical analysis treats Claude Fable 5 and Claude Mythos 5 as two configurations of the same underlying Mythos-class model differentiated solely by safety constraint profiles. Fable 5 targets general availability with classifier-based safeguards that redirect high-risk queries to Opus 4.8 Parameter count is not disclosed, but performance characteristics — especially sustained autonomous multi-hour tasks across millions of tokens — suggest an architecture optimized for long-horizon reasoning and agentic workflows. The article situates Fable 5 as Anthropic's first general-release Mythos-class model, with

A June 10, 2026 developer guide documents the June 9, 2026 release of Claude Fable 5, described as the first Mythos-class model made generally available. Fable 5 posts 80.3% on SWE-Bench Pro with the next-best model reportedly 11 points behind, and is framed as built for long-running, asynchronous tasks such as large-s The article situates Fable 5 one tier above Opus in Anthropic's lineup (Haiku, Sonnet, Opus, Mythos-class), and traces every figure back to Anthropic's launch materials. It is aimed at developers putting Fable 5 into production, covering what the model is, how it benchmarks, what it costs, and integration steps. Benchm