Featured fable-5-1 mythos-5-1 opus-5-5 sonnet-5-5 haiku-5-5 claude-code models

Claude Fable 5.1 and the Claude 5.5 Family: What Changed for Claude Code

Fable 5.1, Mythos 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5 explained: launch dates, pricing, minimum Claude Code versions, alias switches, the Opus default, and effort defaults.

October 21, 2026 • 13 min read • By Claude World

Between September 1 and October 7, 2026, Anthropic replaced the model at every tier of the Claude lineup. Claude Fable 5.1 and Claude Mythos 5.1 came first. Claude Opus 5.5, which Anthropic calls “the first model in our new Claude 5.5 family”, followed on September 22, then Claude Sonnet 5.5 and Claude Haiku 5.5. Claude Code added each model on its launch day.

Two of those Claude Code changes reach you without touching a setting: on the Anthropic API, the opus, sonnet, haiku and fable aliases now resolve to the new models, and the default model on Pro, Max, Team and Enterprise is now Opus 5.5. This article collects the dates, prices, versions and defaults in one place, using only what Anthropic has published, and ends with a practical way to choose a model in Claude Code.

For the other Claude Code changes from the same period, see the companion August–October 2026 roundup.

The Launches at a Glance

DateModelAPI IDClaude Code
2026-09-01Claude Fable 5.1claude-fable-5-1v2.1.257 or later
2026-09-01Claude Mythos 5.1claude-mythos-5-1No alias or documentation
2026-09-22Claude Opus 5.5claude-opus-5-5v2.1.280 or later
2026-09-28Claude Sonnet 5.5claude-sonnet-5-5v2.1.284 or later
2026-10-07Claude Haiku 5.5claude-haiku-5-5v2.1.293 or later

All five have a 1M-token context window, 128K tokens of maximum output, and a reliable knowledge cutoff of June 2026. Anthropic’s models overview now lists Fable 5, Opus 5, Opus 4.8, Opus 4.7, Opus 4.6, Opus 4.5, Sonnet 5, Sonnet 4.6 and Haiku 4.5 as legacy models that are still available.

Where Each Model Fits

Anthropic’s models overview gives the shortest answer: “If you’re unsure which model to use, start with Claude Opus 5.5 for most workloads.” The rest of the lineup is positioned around that default.

Claude Fable 5.1: the most capable generally available model

The models overview says: “Use Claude Fable 5.1 for demanding reasoning and long-horizon agentic work, or when your evals on Claude Opus 5.5 at higher effort still fall short.” The Claude Code docs put it this way: “Claude Fable 5.1 and Claude Fable 5 are the most capable models in Claude Code, suited to tasks larger than a single sitting.”

Headline claims from Anthropic’s announcement:

“They’re the world’s most advanced models for coding and knowledge work—and their research capabilities offer an early glimpse of how AI models will contribute to scientific progress.”

“Fable 5.1 will cost an estimated 25% less than Fable 5 for typical workloads, wherever usage is billed by token.”

“Claude Code users can expect an average of around 60% fewer interventions per session from our cyber safeguards, relative to the previous safeguards on Fable 5.”

The list price is unchanged from Fable 5 at $10 / $50 per million input / output tokens. Cache reads cost $0.25 per million tokens, 0.025x the base input price.

Claude Opus 5.5: the new default

The models overview describes Opus 5.5 as “For long-running agentic coding and knowledge work”. Headline claims from Anthropic’s announcement:

“It performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5.”

“On GDPval-AA v2.1, a test of real-world work across 44 occupations, Opus 5.5 scores 1846 Elo, ahead of Fable 5.1 and Opus 5.”

“Opus 5.5 also generates output more than 30% faster than Opus 5.”

On price, the announcement says: “Input and output tokens are $4 and $20 per million, 20% less than Opus 5. Cache reads (which make up the majority of agentic and coding work costs) are $0.20 per million tokens, 60% less than Opus 5.” Fast mode is available as well: “Fast mode for Opus 5.5 is also available in Claude Code and the Claude Platform with up to 2.5x speed. It costs $8 per million input tokens and $40 per million output tokens.” Fast mode for Opus 5.5 left research preview on the Claude API on October 15, and Claude Code v2.1.304 removed the “(research preview)” label from the /fast dialog on Opus 5.5. With the launch, Anthropic also said it is “increasing five-hour usage limits on Pro, Max, Team, and seat-based Enterprise plans”.

Claude Sonnet 5.5: speed and intelligence

The models overview calls Sonnet 5.5 “The best combination of speed and intelligence”, and the announcement describes it as “a faster, lower-cost complement to Claude Opus 5.5”. Headline claims from Anthropic’s announcement:

“It’s a clear upgrade over Claude Sonnet 5, runs 30%+ faster, and costs up to 30% less for most work.”

“Sonnet 5.5 scores 70.6% on Terminal-Bench 4.0, an agentic coding evaluation, compared to Sonnet 5’s 10.3%.”

“Because its cybersecurity capabilities are comparable to Opus 5’s, it’s the first Sonnet model to launch with cyber safeguards and fallbacks like those we’ve developed for our most capable models.”

Sonnet 5.5 launched with cache reads at $0.20 per million tokens; on October 7, Anthropic lowered that to $0.10. One detail is specific to Claude Code: Sonnet 5.5’s thinking blocks are bound to the account that produced them, and the announcement points anyone who moves conversations between accounts, “including switching accounts mid-session in Claude Code”, to Anthropic’s preserved thinking docs.

Claude Haiku 5.5: the small model, now with effort

The models overview positions Haiku 5.5 “For high-volume, latency-sensitive tasks such as classification, extraction, and routing”, and the announcement adds: “It pairs well with Opus 5.5 and Sonnet 5.5 as a subagent on coding work.” Headline claims from Anthropic’s announcement:

“Introducing Claude Haiku 5.5: the cheapest, fastest, and most capable small model we’ve ever released.”

“Haiku 5.5 is available at a much lower price than Haiku 4.5. On average, it now costs around 75% less to run.”

“Sonnet 5.5 and Opus 5.5 remain better choices for complex agentic coding tasks like those measured by Terminal-Bench 4.0.”

Two practical details: Haiku 5.5 has two price tiers by prompt length (prompts over 100K tokens cost more), and on its tokenizer “the same text counts as approximately 30% more tokens than on Claude Haiku 4.5”.

Prices, Limits, and Default Effort

ModelInput / output per MTokCache reads per MTokContext / max outputDefault effort (API)Default effort (Claude Code)
Fable 5.1$10 / $50$0.251M / 128Khighhigh
Mythos 5.1$10 / $50$0.251M / 128KhighNot in the Claude Code docs
Opus 5.5$4 / $20$0.201M / 128Kmediummedium
Sonnet 5.5$2 / $10$0.10 ($0.20 until October 7)1M / 128Khighmedium
Haiku 5.5$0.10 / $0.50 for prompts up to 100K tokens; $0.50 / $2.50 above 100K$0.01 / $0.05 by the same tiers1M / 128Kmediummedium

Batch API requests are 50% off on every model in the table. On the Message Batches API, Opus 5.5, Sonnet 5.5 and Haiku 5.5 can return up to 300K output tokens with the output-300k-2026-03-24 beta header.

Mythos 5.1: Fable 5.1 With Different Safeguards

The structure is the same one we described for Fable 5 and Mythos 5: one model, two access paths. Anthropic’s announcement is explicit:

“Claude Fable 5.1 and Claude Mythos 5.1 are the same model, but with different levels of safeguards. Fable 5.1 is generally available, while Mythos 5.1 is available only through our trusted access programs; its safeguards are specifically designed to support work in cybersecurity and the life sciences.”

“Claude Mythos 5.1 is identical to Fable 5.1, but it offers more permissive safeguards for vetted individuals and organizations whose work is affected by the cybersecurity and life sciences restrictions outlined above.”

The trusted access programs are:

  • Cyber Verification Program (CVP). On October 6, Anthropic expanded it into Defense Access, Red Team Access and Specialized Access tiers: “Each tier includes access to our most capable models, including Claude Opus 5.5, Claude Sonnet 5.5, Claude Mythos 5.1, and new models moving forward.”
  • Life Sciences Verification Program (LSVP). Launched on September 17: “Standard Use grants apply to Mythos 5.1, Opus 5, and Sonnet 5 today, and to future models as they launch.”

At launch, the announcement said of Mythos 5.1: “Currently, it is only available to a set of US organizations, though we’re coordinating with the US government to expand access to a broader set of domestic and international partners as quickly as possible.” Mythos 5.1 also powers Claude Security, Anthropic’s product that scans codebases for vulnerabilities and suggests patches for human review. Its specifications and pricing match Fable 5.1.

In Claude Code there is no mythos alias, and neither the model configuration docs nor the CHANGELOG mention Mythos 5.1. For everyone outside those programs, Fable 5.1 is the same model; nothing smarter sits behind the gate.

What Changed Inside Claude Code

Minimum versions and alias switches

ModelMinimum versionWhat changed in that version
Fable 5.1v2.1.257fable resolves to Fable 5.1, except in Claude apps gateway sessions, where fable and best still resolve to Fable 5
Opus 5.5v2.1.280opus resolves to Opus 5.5 on the Anthropic API, Claude Platform on AWS, Amazon Bedrock and Google Cloud’s Agent Platform
Sonnet 5.5v2.1.284sonnet resolves to Sonnet 5.5 on the Anthropic API
Haiku 5.5v2.1.293haiku resolves to Haiku 5.5 on the Anthropic API

If you use the Anthropic API directly and picked Fable 5 in /model before v2.1.257, Claude Code changed the saved claude-fable-5 value in your user settings to the fable alias the first time v2.1.257 or later ran. The docs note: “The startup model line shows (auto-updated) once.”

Aliases still depend on your provider

Provideropussonnethaiku
Anthropic APIOpus 5.5Sonnet 5.5Haiku 5.5
Claude Platform on AWSOpus 5.5Sonnet 4.6Haiku 4.5
Amazon Bedrock, Google Cloud’s Agent PlatformOpus 5.5Sonnet 4.5Haiku 4.5
Microsoft FoundryOpus 4.6Sonnet 4.5Haiku 4.5

Where an alias still resolves to an older model, select the newer model by its full name, or pin it with ANTHROPIC_DEFAULT_OPUS_MODEL or ANTHROPIC_DEFAULT_SONNET_MODEL.

Opus is now the default model

v2.1.280 “Changed the default model on Pro and Team Standard plans from Sonnet to Opus, matching Max, Team Premium, and Enterprise”. With no model configured, default now resolves to Opus 5.5 on Pro, Max, Team, Enterprise and the Anthropic API, and on Claude Platform on AWS, Amazon Bedrock and Google Cloud’s Agent Platform. Microsoft Foundry defaults to Sonnet 4.5. Fable is still opt-in: “Neither Fable model is the account-type default on any plan or provider.”

Default effort: medium on the 5.5 models

In Claude Code, the default effort is “high on every model that supports effort, except that Opus 5.5, Sonnet 5.5, and Haiku 5.5 default to medium, Opus 4.7 defaults to xhigh”. Two consequences:

  • Sonnet 5.5 starts lower in Claude Code than on the API. From the announcement: “In Claude Code and our apps, the default effort is set to Medium, while the Claude Platform defaults to High.”
  • Opus 5.5 at medium is not a downgrade. From the Claude Code docs: “In Anthropic’s testing, Opus 5.5 at medium matches or exceeds Opus 5 at high on coding and knowledge-work evaluations.”

v2.1.280 also “Changed an effort level saved before /effort became per-model to no longer apply to newly released models such as Opus 5.5; they start at their default until you pick a level”. Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5 all support low, medium, high, xhigh and max.

Haiku 5.5 brings effort to the Haiku tier

“Haiku 5.5 is our first Haiku-class model to come with an adjustable effort setting.” In Claude Code it defaults to medium, and the docs add: “Sessions auto-compact at about 967K tokens by default.” The catch is the provider table above. haiku means Haiku 5.5 only on the Anthropic API; on Claude Platform on AWS, Amazon Bedrock, Google Cloud’s Agent Platform and Microsoft Foundry it still means Haiku 4.5, which has no effort setting.

Sonnet 4.5 Is Deprecated and Retires November 30, 2026

The one model deprecation in this period: “On September 30, 2026, Anthropic notified developers using Claude Sonnet 4.5 of its upcoming retirement on the Claude API.” claude-sonnet-4-5-20250929 retires on November 30, 2026, and the recommended replacement is claude-sonnet-5-5.

The dates on Anthropic’s deprecations page apply to Anthropic-operated platforms: the Claude API, Claude Platform on AWS and Microsoft Foundry. “Partner-operated platforms (Amazon Bedrock and Google Cloud) set their own retirement schedules”, so dates there can differ.

This matters in Claude Code because sonnet still resolves to Sonnet 4.5 on Amazon Bedrock, Google Cloud’s Agent Platform and Microsoft Foundry, and Foundry’s default model is Sonnet 4.5. Before November 30:

  • Search your settings, subagent definitions, fallbackModel chains and scripts for claude-sonnet-4-5.
  • On Microsoft Foundry, select a newer model explicitly or pin one with ANTHROPIC_DEFAULT_SONNET_MODEL.
  • On Amazon Bedrock and Google Cloud, check your provider’s own retirement schedule.

The new models come with longer commitments: Anthropic will not retire Fable 5.1 or Mythos 5.1 sooner than September 1, 2027, Opus 5.5 sooner than September 22, 2027, Sonnet 5.5 sooner than September 28, 2027, or Haiku 5.5 sooner than October 7, 2027.

How to Choose in Claude Code

Start from the default and move up or down per task:

WorkModelEffortHow
Most coding and knowledge workOpus 5.5, the defaultmedium, the defaultNothing to change
Bug fixes where verification and edge cases matterOpus 5.5high/effort high
Long, ambiguous tasks, or work where Opus 5.5 at higher effort still falls shortFable 5.1high, the default/model fable
Everyday tasks where speed and cost matter moreSonnet 5.5medium, the default/model sonnet
High-volume subagent work: scanning, classification, extractionHaiku 5.5medium, the defaultmodel: haiku in the subagent

Inside a session:

/model                    # open the picker: Enter saves a default, s switches for this session only
/model opus               # Opus 5.5 on the Anthropic API (v2.1.280+)
/model sonnet             # Sonnet 5.5 on the Anthropic API (v2.1.284+)
/model fable              # Fable 5.1 (v2.1.257+); never the default
/model claude-haiku-5-5   # Haiku 5.5 by full model ID (v2.1.293+)
/effort high              # saved per model; press s in the slider for this session only
/effort status            # show the effort level in effect

From the shell:

claude --version                                  # v2.1.293 or later covers all four aliases
claude --model fable                              # one session on Fable 5.1
claude --model opus --effort high                 # Opus 5.5 at high, for this session only
claude --model claude-sonnet-5-5                  # pin an exact model ID
claude -p "List the failing tests" --model haiku  # quick headless run on Haiku 5.5

To keep a subagent on the small model no matter what the session uses, pin it in the agent definition:

---
name: log-scanner
description: Reads test and build logs and reports only the relevant failures.
tools: Read, Grep, Glob
model: haiku
---

Report each failing test with its file and line. Never modify files.

Three habits that prevent surprises:

  1. /model also moves subagents that inherit the session model. The docs: “To keep a custom subagent on a smaller model, set model in its definition.”
  2. Check how Fable bills before you automate it. “Depending on your plan and seat tier, Fable usage can bill to usage credits instead of drawing on your plan’s included limits.” Interactive sessions ask for consent first; in a claude -p run, Claude Code bills it without asking.
  3. Expect safety fallbacks in security and biology work. For Opus 5.5, Anthropic says “most cybersecurity tasks will be re-routed to Opus 4.8”. Claude Code shows a notice in the transcript when a flagged request re-runs on a fallback model, and /model takes you back.

Sources

Keep going: the August–October 2026 Claude Code roundup covers permission defaults, ultracode and the other changes from the same releases, and Session 26: Effort Levels, Ultrathink & Ultracode is the hands-on lesson on effort.