Anthropic Model Names Explained: Fable, Mythos, Opus, Sonnet

Last updated: October 9, 2026 - Haiku 5.5 (released October 7) replaces Haiku 4.5 in the current lineup; table, model IDs, prices and FAQ re-verified against Anthropic's models overview.

Anthropic names Claude models by tier and version: Haiku is the fastest and cheapest, Sonnet balances speed and intelligence, and Opus handles long-running agentic coding and knowledge work. Fable is a newer named tier above Opus, and Mythos is the class it belongs to. Anthropic describes Fable 5 as "a Mythos-class model that we've made safe for general use", while Mythos models themselves have limited availability. As of October 9, 2026, the current lineup is Fable 5.1, Opus 5.5, Sonnet 5.5, and Haiku 5.5.

Official Sources#

The old ladder#

For most of Claude's public history, the naming system fit in a README comment: Haiku, Sonnet, and Opus, ordered by capability and cost, each with a version number. Claude 3 Opus, Claude 3.5 Sonnet, and Claude 4 Opus told you the capability class and how recent the model was.

What Mythos and Fable changed#

With Claude Fable 5, launched June 9, 2026, Anthropic introduced a named class above Opus. Per the announcement, Fable 5 is a Mythos-class model made safe for general use, with safeguards that send queries on some topics to Claude Opus 4.8 instead (Anthropic says they trigger in less than 5% of sessions on average). Claude Mythos 5 is the same underlying model with the safeguards lifted in some areas, initially deployed through Project Glasswing for cyberdefenders. Anthropic's pricing page still lists Mythos models as limited availability.

So the model name and the class name are different things. claude-fable-5-1 does not contain the word Mythos, because Mythos is the capability class and Fable is the generally available flagship inside it. That separation lets Anthropic ship a general-use model and a restricted-access sibling from the same underlying weights.

Current model tiers#

From Anthropic's models overview, checked October 9, 2026:

Older models such as Fable 5, Opus 5, Opus 4.8, and Sonnet 5 remain available as legacy models. Anthropic's guidance is to start with Opus 5.5 for most workloads, use Fable 5.1 for demanding reasoning and long-horizon agentic work or when your evals on Opus 5.5 at higher effort still fall short, and pick a cheaper tier when evals show quality holds. The tier also changes the default effort setting, so see Fable 5 effort levels explained before comparing runs across models.

What the naming signals#

Names are roadmap signals, and the June questions have mostly answered themselves. There is a Fable 5.1, so named models do get point releases. There are Mythos 5 and Mythos 5.1, so the class holds more than one model. Opus did not stop at 4.x: it moved to Opus 5 and Opus 5.5, and it is now priced below the Opus 4.x models it replaced ($4 input and $20 output for Opus 5.5 versus $5 and $25). The takeaway is that price bands and model IDs are more reliable than the tier names themselves, so read the models overview instead of inferring capability from a name.

How to choose now#

• Start with the tier Anthropic recommends for your kind of work and move up only when your evals fail. See Fable 5 vs Opus 4.8: when to use which and the Opus 5 vs Opus 4.8 vs Fable 5 comparison.
• Model the cost per completed task, not per token. See Fable 5 production cost modeling and the cost-per-task analysis.
• Pin the exact model ID in production. Anthropic states every Claude model ID is a pinned snapshot, so you control when upgrades happen.
• If you are moving from older Claude models, the migration checklist for Fable 5 lists the breaking changes.
• For the whole family on one page, see the Claude tool page; for how it stacks up against OpenAI's lineup, see Claude vs ChatGPT.

FAQ#

What is Claude Mythos?#

Mythos is Anthropic's top capability class, above Opus. Claude Fable 5 is described as a Mythos-class model made safe for general use, while Mythos 5 and Mythos 5.1 have safeguards lifted in some areas and are limited-availability models.

What is the difference between Fable and Mythos?#

Per the launch announcement, Mythos 5 is the same underlying model as Fable 5 with the safeguards lifted in some areas. Fable is the generally available model, and Mythos is offered to a small group of cyberdefenders and infrastructure providers under trusted access.

Which Claude model should I use for coding?#

Anthropic's models overview describes Opus 5.5 as built for long-running agentic coding and recommends starting there for most workloads. Move to Fable 5.1 for the most demanding long-horizon work, or down to Sonnet 5.5 for a faster and cheaper tier once your evals show quality holds.

What model IDs should I use in API calls?#

Use the IDs from the models overview: claude-fable-5-1, claude-opus-5-5, claude-sonnet-5-5, and claude-haiku-5-5. Every Claude model ID is a pinned snapshot, so pin the exact ID in production.

How much does each Claude tier cost?#

Per million tokens: Fable 5.1 is $10 input and $50 output, Opus 5.5 is $4 and $20, Sonnet 5.5 is $2 and $10, and Haiku 5.5 starts at $0.10 and $0.50 (listed as "from" prices; the rate rises for prompts over 100,000 tokens, see the Haiku 5.5 guide). Prompt cache reads and batch requests are cheaper, per Anthropic's pricing page.

Continue Reading#

• Claude Haiku 5.5: Pricing, Migration Changes, and When to Use It - the newest tier and its pricing cliff
• Fable 5 Effort Levels Explained - how effort changes token use and cost across models
• Claude Fable 5 vs Opus 4.8: When to Use Which - the tier decision in detail
• Claude Code Fast Mode: When 2.5x Speed Is Worth 2x Price - the other lever on speed and cost
• Apple's LanguageModel Protocol: Xcode 27 Just Made Model Lock-In Optional