TL;DR. A vendor page gathering Anthropic's base models. Per-model strengths, benchmarks, and operational constraints are backed by primary sources in each standalone piece below; here we only classify which model takes which job.
What this slot is
The set of base models Anthropic ships for general use. It is one vendor column in the models slot that coding agents, frameworks, and skills ride on — and even within one vendor the job splits per model: everyday vs heavy autonomous runs, generalist vs specialized.
Lineup — two axes, and the gap between them closed (as of 2026-08)
Anthropic's base models split into a top-intelligence axis (Fable/Mythos) and a working axis (Opus/Sonnet). What matters is that the gap between them nearly vanished in July: Claude Opus 5 (2026-07-24) came in at SWE-bench Verified 96.0% and SWE-bench Pro 79.2%, putting it under a point behind Fable 5 (80.0) on SWE-bench Pro at half the price ($5/$25 vs $10/$50). On Anthropic's own agentic-coding eval, Opus 5 leads Fable 5.
So the adoption question for this vendor is no longer "which is strongest" but "is there still a job that justifies the Fable 5 premium?" Fold in the access constraints below and that answer narrows further.
Models
Each model's strengths and weaknesses are sourced in its standalone piece.
- Claude Fable 5 — strong on long-horizon autonomy, coding, vision, and low hallucination; weak on non-top science reasoning, classifier over-blocking, 2× price, and operational constraints. Access was cut by government directive on 2026-06-12 and restored globally on 07-01 (Mythos 5 stays limited to a set of US organizations).
- Claude Opus 5 (2026-07-24,
claude-opus-5) — no standalone piece yet. SWE-bench Verified 96.0%, Pro 79.2%, 1M context, unchanged $5/$25. Currently this vendor's default on price-performance.