Claude

Claude Opus 5.5 & Fable 5.1 — Anthropic

Anthropic's precision agent for complex tasks. Opus 5.5 is the new flagship — Anthropic says it performs at Fable 5.1's level on most work, at $4 / $20 and 40% cheaper to run than Opus 5. The catch: thinking can't be switched off anymore, and a few API calls that worked on Opus 5 now return 400. Fable 5.1 stays the ceiling for autonomous runs that last days. The 1M context still reads your entire codebase and remembers page 847.

Opus 5.5 Released

Sep 22, 2026

Context

1M tokens

Starting Price

$20/mo (Pro)

Max Output

128K tokens

01

Overview

Claude is Anthropic's flagship AI assistant, designed with a focus on being helpful, harmless, and honest. It's not the flashiest name in AI, but it's the model that developers and enterprises keep coming back to when the work actually matters.

Anthropic refreshed the whole lineup in September: Fable 5.1 (Sep 1), Opus 5.5 (Sep 22), Sonnet 5.5 (Sep 28). Claude Opus 5.5 is the one that matters for most people. It's cheaper than the model it replaces — $4 / $20 per million tokens, down from Opus 5's $5 / $25 — and Anthropic says it costs 40% less to run than Opus 5 with output more than 30% faster. Reportedly it's also the Claude Code default on Pro, Max, Team, Enterprise, and the API.

The headline claim is that Opus 5.5 "performs at the level of Claude Fable 5.1 on most work." If that holds for your workload, the top tier just became a niche product. Test it on your own tasks before you believe it — but the price drop alone makes Opus 5.5 the default pick.

Anthropic's own numbers: 66.4% on Terminal-Bench 4.0, 54.4% on FrontierCode v1.1, 57.8% on CursorBench 4.0, 81.8% on OSWorld 2.1 computer use, and 1,846 Elo on GDPval-AA v2.1 knowledge work. Vendor benchmarks, vendor-picked — useful for direction, not gospel.

Now the bad news. Opus 5.5 ships breaking API changes. Thinking is always on — thinking: {type: "disabled"} and budget_tokens both return a 400, and effort is the only dial left. Default effort dropped from high to medium (Anthropic says 5.5 at medium matches or beats Opus 5 at high). Forced tool calls, non-default sampling parameters, and assistant prefill are all rejected. If you're migrating an app, read the Limitations section before you flip the model ID.

Claude Fable 5.1 is still the ceiling and still costs $10 / $50 — now 2.5x Opus 5.5's per-token price, with cache reads 75% cheaper than Fable 5's. It remains the pick for long-horizon agents that run autonomously for days, and it's never a default anywhere. Above it sits Claude Mythos 5.1: same specs and price as Fable 5.1, invitation-only through Project Glasswing.

Claude Sonnet 5.5 rounds out the lineup at $2 / $10 with the same 1M context and 128K output. Opus 5, Fable 5, and Sonnet 5 are now legacy but still available.

Access the free tier at claude.ai (Sonnet and Haiku, no Claude Code) or upgrade to Pro ($20/mo) for Opus 5.5 plus capped Fable access. On the API it's claude-opus-5-5.

02

Key Capabilities

Fable-Level Work at Opus Prices (Opus 5.5)

Per Anthropic, Opus 5.5 performs at the level of Fable 5.1 on most work — at $4 / $20, 40% cheaper to run than Opus 5, with output more than 30% faster

Benchmarks (Opus 5.5, per Anthropic)

66.4% Terminal-Bench 4.0, 54.4% FrontierCode v1.1, 57.8% CursorBench 4.0, 81.8% OSWorld 2.1, 1,846 Elo on GDPval-AA v2.1 — vendor numbers, so run your own evals

Effort Control

Five levels — low, medium, high, xhigh, max. Same per-token rate at every level. On Opus 5.5 it's the only reasoning dial, and the default dropped from high to medium — Anthropic says 5.5 at medium matches or beats Opus 5 at high. Re-run your effort sweep instead of carrying old settings over.

Fast Mode (Opus 5.5)

Faster output for double the rate ($8 / $40 per million) — cheaper than Opus 5's fast mode was. Also available in Claude Code

High-Resolution Vision

High-res image support up to 2576px / 3.75MP. Big jump for design reviews, dense diagrams, and document analysis

Adaptive Thinking

Automatically adjusts reasoning depth — fast for simple tasks, deep for complex ones. On Opus 5.5 it's always on; you steer it with effort, not by switching it off

Multi-Agent Coordination

Runs writer-verifier patterns cleanly and rarely lets subagents overwrite each other. Since Opus 5, Claude delegates more readily than 4.8 did, so cap spawn counts if you're cost-sensitive

Computer Use

Navigates interfaces, fills forms, moves data across applications — Opus 5.5 scores 81.8% on OSWorld 2.1, per Anthropic

1M Token Context (Standard)

1M context on Opus 5.5 and Sonnet 5.5. Process entire codebases, legal documents, or research papers in one conversation

128K Max Output (300K Beta)

128K standard, or up to 300K tokens on the Batch API via beta header (output-300k-2026-03-24)

Compaction

Automatic server-side context summarization when conversations approach the limit

Cheaper Prompt Caching

Cache reads cost $0.20 per million on Opus 5.5 and $0.25 on Fable 5.1 — 75% cheaper than Fable 5. Opus 5 already dropped the minimum cacheable prompt to 512 tokens, down from 1024 on Opus 4.8

Long-Horizon Autonomy (Fable 5.1)

Fable 5.1 remains the pick for agents that run unattended for days. In Claude Code it's /model fable. Single requests can take many minutes — plan your timeouts and progress UI accordingly

Long-Context Coherence

The full 1M window bills at standard rates — a 900K-token request costs the same per token as a 9K one. No long-context surcharge, unlike GPT-6 and Grok.

03

Pricing

Tier	Price	Details
Free	$0	Sonnet and Haiku. No Claude Code. Projects, Artifacts, app connectors, and designs/decks/docs in any chat. Tighter limits.
Pro	$20/mo ($17/mo annual)	~5x free tier messages. Opus 5.5, plus Fable capped at 50% of weekly limits. Higher five-hour limits since Sep 22.
Max 5x	$100/mo	25x free tier capacity. Opus 5.5, plus Fable capped at 50% of weekly limits.
Max 20x	$200/mo	100x free tier capacity. Opus 5.5 for daily work, Fable 5.1 for the heaviest runs.
Team	$20 standard / $100 premium per seat/mo	Collaboration features, admin controls, Opus 5.5 access. Fable capped at 50% of weekly limits.
Enterprise	Custom	SSO, audit logging, compliance APIs, enhanced context.
API — Opus 5.5	$4 / $20 per 1M tokens	The flagship, down from Opus 5's $5 / $25. Cache reads $0.20. Fast Mode $8 / $40. Batch API at 50% off.
API — Fable 5.1	$10 / $50 per 1M tokens	Top tier, for multi-day autonomous agents. Cache reads $0.25 — 75% cheaper than Fable 5.
API — Sonnet 5.5	$2 / $10 per 1M tokens	Mid-tier at a fifth of Fable's price. 1M context, 128K output. Legacy Sonnet 5 is also $2 / $10 — the scheduled rise to $3 / $15 was called off.
API — Haiku 4.5	$1 / $5 per 1M tokens	Fastest, most affordable. 200K context. Haiku 5.5 is reportedly due "in the coming weeks" (not out yet), and 4.5 reportedly retires no sooner than Oct 15, 2026.

04

Best Use Cases

Complex software development and large codebase management

Long, sustained coding sessions and agentic coding tasks

Multi-application workflow automation via computer use

Enterprise document analysis — legal, financial, medical

Research tasks requiring long-context coherence across massive documents

Building AI-powered applications via the API

Tasks requiring safety, compliance, and governance

05

Pro Prompts

Codebase Analysis

prompt.txt

Debug this large codebase and suggest refactors. Prioritize by impact-to-effort ratio. Flag any security concerns you spot.

Scope Discipline (Opus 5.5)

prompt.txt

Deliver exactly what I asked for, at the scope I intended. Make routine judgment calls yourself; check in only when two readings would lead to materially different work. If you think the ask is wrong, say so in one sentence and keep going. Don't add helpers, abstractions, or error handling for cases that can't happen. Task: [task].

Max Effort (Fable 5.1 / Opus 5.5)

prompt.txt

Set effort to max. This is the kind of problem that justifies the cost: [hard problem]. Take the time. Verify every step. I'd rather wait 5 minutes for a right answer than 30 seconds for a wrong one.

Document Deep Dive

prompt.txt

Analyze this 200-page contract. Extract every obligation, deadline, and liability clause. Flag anything that deviates from standard commercial terms.

Computer Use

prompt.txt

Navigate to [application], export the Q4 report as CSV, clean the data, and create a summary dashboard in Google Sheets.

06

Limitations

Higher latency at high/max effort levels, especially on complex reasoning tasks

You can't turn thinking off on Opus 5.5. thinking: {type: "disabled"} and budget_tokens both return a 400 — effort is the only dial. Default effort is now medium (Opus 5 was high), so re-run your effort sweep instead of trusting old settings. Per Anthropic's migration guide; reportedly the same on Fable 5.1 and Sonnet 5.5.

Forced tool calls are gone: tool_choice any or tool returns a 400 on Opus 5.5. Use auto plus strict tool use or structured outputs. Non-default temperature/top_p/top_k and assistant prefill are rejected too.

Text Opus 5.5 writes between tool calls now arrives inside thinking blocks, empty at the default display — a UI that streamed it goes quiet unless you set display to "updates" or "summarized"

Opus 5.5 thinks on every request, and max_tokens caps thinking plus response together — budgets tuned for older models can truncate mid-answer

Spatial reasoning remains a common weakness

Since Opus 5, Claude reaches for subagents more eagerly than 4.8 did — great for coverage, expensive if you don't cap spawn counts

Free tier does NOT include Opus 5.5, Fable, or Claude Code — Sonnet and Haiku only

Fable 5.1 is 2.5x the per-token cost of Opus 5.5 ($10/$50 vs $4/$20) — and if Anthropic's "Fable-level on most work" claim holds, that premium buys you little outside multi-day agents. On the apps, Fable is capped at 50% of your weekly limits.

Claude's frontier models ship cybersecurity and biology safeguards that can refuse a request outright. Legitimate security and life-sciences work does hit false positives.

Claude Mythos 5.1 (same specs and price as Fable 5.1) is invitation-only through Project Glasswing — you can't just sign up for it

Writes longer answers and longer files than previous models. Lowering effort doesn't reliably shorten them — you have to ask for concision explicitly.

No native image generation — text and code output only

07

Multimodal Capabilities

Vision (Up to 3.75MP)

High-res charts, graphs, technical diagrams
Reports, screenshots, dense documents
Max image input: 2576px / 3.75MP

Computer Use

Desktop software operation
GUI navigation and form filling

File Uploads

PDFs and documents
Spreadsheets and presentations

Output

Text and code only
No native image, audio, or video generation

08

Recent Updates

Sep 28, 2026

Claude Sonnet 5.5 released — $2/$10, 1M context, 128K output. Sonnet 5 moves to legacy, and stays at $2/$10: the scheduled increase to $3/$15 was cancelled.

Sep 22, 2026

Claude Opus 5.5 released — $4/$20 (down from $5/$25), Fable 5.1-level on most work per Anthropic, default effort now medium. Breaking API changes: thinking always on, forced tool_choice and prefill rejected. Higher five-hour limits on Pro, Max, Team, and Enterprise the same day.

Sep 16, 2026

Chat and Cowork merged into one experience on Pro and Max. Designs, decks, and docs can now be made in any conversation on every plan, Free included.

Sep 1, 2026

Claude Fable 5.1 released — still $10/$50, cache reads cut 75% to $0.25. Its sibling Mythos 5.1 has the same specs and price, invitation-only via Project Glasswing.

Aug 3, 2026

Claude Code v2.1.221 — Focus view collapses tool activity into compact summaries and declutters long sessions

Jul 24, 2026

Claude Opus 5 generally available — 96.0% SWE-bench Verified, self-verification, 1M context standard, $5/$25 pricing. Default across claude.ai, Claude Code, and Cowork, and live in GitHub Copilot the same day.

Jun 9, 2026

Claude Fable 5 generally available — the top tier, $10/$50 API pricing. Mythos 5 deployed the same day via "Project Glasswing" for infrastructure providers and vetted cybersecurity researchers.

May 28, 2026

Opus 4.8 launched — Dynamic Workflows (parallel subagents), effort control, cheaper Fast Mode. Superseded by Opus 5 at identical pricing.

Apr 16, 2026

Claude Opus 4.7 generally available — step-change agentic coding, high-res vision (3.75MP)

Milestone

Available across all Claude products and the API, plus Amazon Bedrock, Google Vertex AI, GitHub Copilot, and Microsoft Foundry