Note on sources: The draft you shared contained several inaccurate claims (wrong release dates, invented pricing, and unverified training details). This page replaces those with figures pulled from Anthropic's official platform docs and product announcements as of September 3, 2026. I skipped adding logos/screenshots since those are copyrighted assets I can't reproduce — the icon above is just an emoji placeholder you can swap for your own brand image.

Architecture Overview — How the Lineup Is Structured

Architecture blueprint

Blueprint view — illustrative stand-in for the tier structure below.

Anthropic organizes Claude into tiers by capability/cost, not a single flagship:

Claude Haiku 4.5 — fastest, cheapest, smallest context window (200K tokens). Good for high-volume, low-latency tasks.
Claude Sonnet 5 — the default balance of speed and intelligence. 1M token context.
Claude Opus 5 — built for complex agentic coding and enterprise workloads; step-change gains over Opus 4.8 in deep reasoning, long-horizon agentic tasks, and test-time compute scaling.
Claude Fable 5.1 — sits above Opus 5 for demanding reasoning and long-horizon agentic work, or when Opus 5 at higher effort still falls short. Anthropic's guidance is to start with Opus 5 for most workloads and reach for Fable 5.1 only when needed.
Claude Mythos 5.1 — same underlying capability as Fable 5.1, but gated to approved organizations only (Project Glasswing), with fewer safety guardrails for sensitive domains like biology, cybersecurity, and LLM R&D.

All four generally-available tiers above Haiku share a 1M token context window and 128K max output.

Model Comparison Table

Model	Context Window	Max Output	Input Price	Output Price	Best Used For
Claude Haiku 4.5	200K tokens	—	$1 / MTok	$5 / MTok	High-volume, low-latency tasks
Claude Sonnet 5	1M tokens	128K tokens	$2 / MTok	$10 / MTok	Everyday coding, balanced speed/intelligence
Claude Opus 5	1M tokens	128K tokens	$5 / MTok	$25 / MTok	Complex agentic coding, enterprise workflows
Claude Fable 5.1	1M tokens	128K tokens	$10 / MTok	$50 / MTok (cache reads at ~$0.25/MTok)	Long-horizon reasoning, multistep research, heavy document/spreadsheet/slide work
Claude Mythos 5.1	1M tokens	128K tokens	Same as Fable 5.1	Same as Fable 5.1	Gated — Project Glasswing partners only

Code-Side Improvements by Tier

Claude Sonnet 5

Balanced pick for day-to-day coding — standard feature work, boilerplate, refactors, and multi-step tool use where latency and cost matter more than maximum reasoning depth.

Claude Opus 5

The biggest jump is in agentic and long-horizon coding: it holds context across larger codebases, plans multi-file changes, and scales its own reasoning effort (via adaptive thinking) when a task is genuinely hard rather than applying a fixed thinking budget.