Tool support requires the appropriate API integration; a supported tool is not automatically active in every chat.
Supported API features and tools
Text and image input; text output
Adaptive thinking with configurable effort
Messages API tool-use workflows with an appropriate integration
Prompt caching with 5-minute and 1-hour write options
Asynchronous Message Batches processing
Before you choose
No native audio, video, or image output; accepting images does not make this an image generator.
Thinking cannot be disabled, and manual enabled/budget_tokens or between_tools configurations are rejected.
Forced tool selection is not supported. Thinking blocks depend on the model and the conversation that produced them.
Fable is not a grant of access to the separate invitation-only Mythos offering.
Adaptive thinking and effort
lowmediumhigh · defaultxhighmax
Adaptive thinking is always on, and high is the default effort. All five levels are supported; use measured quality gains to justify xhigh/max. Thinking consumes output tokens even when hidden. Give max_tokens room for reasoning plus the response, and bound the task with checkpoints rather than an unlimited objective.
Anthropic lists this model as active; released September 1, 2026. Retirement is not sooner than September 1, 2027. This is an availability commitment, not a scheduled retirement.
The displayed cutoff is Anthropic’s reliable knowledge cutoff, published to month precision. A cutoff does not replace checking current facts or supplied evidence.
Per-message effort and turn-scoped system messages are documented beta features, not ordinary app controls.
When migrating from Fable 5, review forced-tool selection and thinking-block validity before reusing an integration.
The overview lists 128,000 maximum output tokens; do not transfer another Claude model’s extended Batch output beta without explicit support.
Put it to work
Start with a more useful prompt.
Original examples from EZ Ai Assist. Adapt these to your task and the features available in your workspace.
Workflow 01
Plan an evidence-led repository investigation
Set review checkpoints before a long engineering task.
Create an investigation plan for the repository problem described below using the supplied architecture and initial evidence. Break the work into bounded stages with a question, evidence to collect, and an exit condition for each. Identify dependencies, likely blind spots, and decisions requiring an owner. Include a stop condition when evidence is insufficient. Do not execute code or assume access to missing systems.
Workflow 02
Reconcile a multi-document research argument
Trace disagreements to assumptions, methods, and evidence quality.
Evaluate the claim below against the supplied research packet. Map the supporting and opposing arguments to source sections, compare methods and assumptions, and identify evidence that cannot be reconciled. Distinguish direct findings from your interpretation. Produce a defensible synthesis and a list of observations that would overturn it. Do not fill gaps with invented citations.
Workflow 03
Audit consistency across business deliverables
Check a report, spreadsheet, and slide outline against one another.
Compare the supplied report, spreadsheet tables, and slide outline for consistency. Track every key metric, date, definition, and recommendation to its source. Identify mismatched units, missing qualifications, and claims that change between formats. Give a correction checklist with the responsible artifact and evidence for each item. Do not edit underlying facts or assume one version is authoritative without instruction.
Developer reference
Anthropic API pricing
These are Anthropic API reference prices, not EZ Ai Assist subscription prices.
Standard rates apply across the full 1M-token context window; no separate long-context surcharge is listed for this model.
Batch processing discounts input and output by 50%. Cache writes and reads have distinct rates; calculate from actual usage rather than applying a cache-read discount to the entire request.
Thinking tokens are billed as output even when their text is hidden. Tool charges and optional US-only inference (1.1× token pricing) can also affect the total.
Cache reads are $0.25 per million tokens, one quarter of Fable 5’s $1 rate. Base input and output remain $10/$50.
Verify current provider pricing before estimating spend. These rates do not describe an EZ Ai Assist subscription.
Common questions
A few things worth knowing.
When should I evaluate Fable 5.1?
Use it for demanding reasoning or long-horizon work when a simpler model does not meet your quality target. Anthropic recommends starting with Opus 5.5 for most workloads; compare using your own tasks and cost constraints.
What changed from Fable 5?
The current model has a newer reliable cutoff, lower cache-read pricing, and additional beta conversation controls. Forced-tool use and thinking-block reuse also require migration checks; it is not merely a new label.
Can I turn off thinking or only hide it?
Thinking is always on. Omitting the visible summary does not stop reasoning or its billing. Use supported effort settings to tune the work, and allow max_tokens to cover both thinking and response text.
Does Fable 5.1 have a 300K output beta?
Its overview lists a 128,000-token maximum. The cited extended-output beta names specific Opus and Sonnet models, not Fable 5.1. Do not assume an unlisted model inherits that limit.
How much cheaper are cache reads than Fable 5?
The listed rate is $0.25 versus $1 per million cache-read tokens. That is not a fourfold discount on an entire request: base input, cache writes, output, and tool charges still need to be counted.
Does Fable access include Mythos?
No. Mythos 5.1 is a separate invitation-only offering. Similar specifications do not imply shared access, and neither provider offering guarantees availability in EZ Ai Assist.
Is September 1, 2027 its end date?
No. It is the earliest retirement date permitted by the published commitment, not an announced shutdown. Review current lifecycle documentation and app availability before relying on a long-lived integration.
Check the source
Official documentation
Specifications and API prices checked on . Example prompts and workflow advice are editorial guidance from EZ Ai Assist.