Anthropic Claude Fable 5.2
Anthropic frontier tier, 20% cheaper than Fable 5.1 ($8/$40) and up to 30% faster: 49.1% FrontierCode 2.0 at Xhigh, 69.9% FrontierSWE v2, 89.3% OSWorld 2.1; at High it matches Opus 5.5 Max on FrontierCode for about half the cost per task. Pitched as orchestrator/advisor for agent teams.
Interactive Token Cost Estimator (Fable 5.2)
Architectural Strengths & Target Workloads
- Long-horizon agentic coding and multi-agent orchestration
- Opus 5.5-level FrontierCode at High effort for ~half the cost per task
- Computer use (computer_toolset_20260801) and vision
Primary Engineering Workloads: Large Migrations & Architecture · Orchestrator / Advisor for Agent Teams · Deep Research.
Frequently Asked Questions (Fable 5.2)
How much does Anthropic Claude Fable 5.2 cost per 1M tokens?
Anthropic Claude Fable 5.2 is priced at $8.00 per 1M uncached input tokens, $0.20 per 1M cached input tokens, and $40.00 per 1M output tokens (claude-fable-5-2; cache reads at 2.5% of input; fast mode $16/$80 (speed: "fast"); 30-day retention required; API default effort high, Claude Code default medium).
What is the context window of Anthropic Claude Fable 5.2?
It supports up to 1M tokens (default and maximum) (1,000,000 tokens) and a maximum generation output of 128,000 tokens.
Claude Fable 5.2 and Mythos 5.2: Anthropic's Frontier Tier Gets 20% Cheaper and Learns to Run Agent Teams
Same 1M-token context, $8/$40 per million tokens instead of $10/$50, and a pitch aimed at orchestrators and advisors. The API also gets stricter about how you replay thinking blocks.
Read Full Technical Dispatch (6 min read) →