8 tools that already support Claude Fable 5.2 (November 2026)

Eight tools support Claude Fable 5.2, going only by their own changelogs and docs: Claude Code, the Claude apps, the Claude API, Google Cloud, Claude Platform on AWS, Amazon Bedrock, Vercel AI Gateway and OpenRouter. Anthropic released the model on November 9, 2026 at $8 per million input tokens and $40 per million output tokens.

Many tools will say "coming soon" this week. A tool is on this list only when its own page says Fable 5.2 works today. For each one you get the model ID, the price, the date of the vendor's note and the exact steps to switch.

TL;DR: If you write code, run /model fable in Claude Code (v2.1.321 or later). If you build on an API, call claude-fable-5-2 on the Claude API, or use Google Cloud or Amazon Bedrock if your company already bills there.

Last checked: November 9, 2026, at 22:20 UTC. We added Amazon Bedrock that evening, after AWS published its Fable 5.2 model card. We add a tool when its own changelog or docs confirm Fable 5.2. Next check: November 10, 2026.

Which tools support Claude Fable 5.2? A quick overview

For coding and daily work:

• Claude Code: best for developers who want Fable 5.2 in the terminal, with one command to switch.
• Claude apps: best for chat, writing and research with no setup, on eligible paid plans.

For building on the API:

• Claude API: best for teams that build their own agents and want every feature first, such as batch and fast mode.
• Google Cloud: best for companies that already bill on Google Cloud and need US or EU regions.
• Claude Platform on AWS: best for teams that pay through their AWS bill. This is not Amazon Bedrock.
• Amazon Bedrock: best for teams already building on Bedrock, with its Guardrails and IAM.

For trying several models behind one key:

• Vercel AI Gateway: best for apps on the Vercel AI SDK that need fallback models.
• OpenRouter: best for testing Fable 5.2 next to other providers' models with one account.

Prices are per million tokens, input then output, as of November 9, 2026, from each vendor's own page.

Claude Fable 5.2 API: the model ID on each platform

The Claude Fable 5.2 model ID is claude-fable-5-2. Anthropic's models overview uses the same ID on the Claude API, Claude Platform on AWS, Google Cloud and Microsoft Foundry. Amazon Bedrock is the exception: its ID is anthropic.claude-fable-5-2, called through us.anthropic.claude-fable-5-2 or global.anthropic.claude-fable-5-2 (AWS model card). Gateways add a prefix and use a dot: anthropic/claude-fable-5.2 on Vercel AI Gateway and OpenRouter.

A first call on the Claude API looks like any other Messages API call. Only the model string changes:

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model": "claude-fable-5-2", "max_tokens": 1024, "messages": [{"role": "user", "content": "Hello"}]}'

The specs are the same on every platform. Fable 5.2 has a 1M token context window, 128K max output and adaptive thinking that is always on, per the Fable 5.2 model page. AWS's Bedrock model card lists the same 1M context and 128K output.

How to use Claude Fable 5.2 in Claude Code

To use Claude Fable 5.2 in Claude Code, update to v2.1.321 or later and select the fable alias. These steps come from the Claude Code model docs.

1. Run claude update. Fable 5.2 needs Claude Code v2.1.321 or later.
2. In a session, run /model fable. Or start with claude --model fable.
3. In Claude apps gateway sessions, run /model claude-fable-5-2 instead, because there the fable alias still points to Fable 5.

You can also set ANTHROPIC_MODEL=fable or put "model": "fable" in your settings file. The docs say no Fable model is the default on any plan, so you have to pick it. If the /model picker shows "Requires usage credits" on the Fable row, Fable usage on your seat bills to usage credits.

1. Claude Code, best for developers who want Fable 5.2 in the terminal

Claude Code is Anthropic's coding agent that runs in your terminal and IDE. It added Fable 5.2 in version 2.1.321, and the fable alias now points to Fable 5.2 by default.

This is the fastest way to put Fable 5.2 on a real codebase. Anthropic's announcement says fast mode is offered in Claude Code too, at a higher price per token. On Pro and Max, a saved Fable 5.1 choice moves to the fable alias the first time you run the new version.

Key features

• fable alias that resolves to Fable 5.2 everywhere except Claude apps gateway sessions
• Switch mid-session with /model fable, or at start with --model fable
• 1M token context window for large codebases
• Optional fast mode
• Picker label "Requires usage credits" when Fable bills to credits

Best for

• Developers on Pro, Max, Team or Enterprise who already use Claude Code daily
• Teams that want to test Fable 5.2 on long refactors before they pay for API traffic

Pricing Fable usage draws on your plan limits or on usage credits, depending on plan and seat tier. Pro and standard Team seats get Fable included through November 16, 2026, up to 50% of weekly limits (as of Nov 2026, Claude support). With an API key, you pay the API rate of $8 / $40 per million tokens.

Pros

• One command to switch, and the change applies right away
• Same model ID as the API, so results carry over to your own agents

Cons

• Needs v2.1.321 or later, so older installs fail until you update
• On some seats Fable bills to usage credits instead of plan limits, which can surprise you on long runs

2. Claude apps, best for chat, writing and research with no setup

The Claude apps are Anthropic's chat apps, including claude.ai. Anthropic's launch offers article says: "To use a Fable model, select it from the model picker." Your default model does not change, so each chat starts on your usual model until you pick Fable.

This suits people who want to try Fable 5.2 on documents, analysis or planning without writing code. The launch offers decide how much you can use before credits apply.

Key features

• Pick Fable 5.2 from the model picker
• Launch offer for Pro and standard Team seats until November 16, 2026
• One to three Fable limit resets for Max, Ultra and premium seats, based on how long you have subscribed
• Usage past the allowance bills to usage credits, if you turned them on

Best for

• Founders and marketers who want Fable 5.2 for research and writing
• Teams on Team plans that want to trial it for a week before they decide on credits

Pricing Included within the launch offer limits on eligible plans: Pro, Max and Ultra, standard and premium Team seats, and premium seats on legacy seat-based Enterprise plans. The article says the offers "aren't available on the Free plan" (as of Nov 2026, Claude support).

Pros

• No setup beyond picking the model
• Max and Ultra resets last until March 8, 2027

Cons

• The Pro and Team allowance ends on November 16, 2026, then usage needs credits
• Anthropic's pages do not state whether the Free plan can use Fable at all

3. Claude API, best for teams building their own agents

The Claude API is Anthropic's own developer platform, where you call models with an API key. Fable 5.2 is listed in the Claude Platform release notes for November 9, 2026, with the ID claude-fable-5-2.

Features land here first. Batch, prompt caching and fast mode are all priced on the pricing page. Anthropic's docs suggest you start most workloads on Opus 5.5 and move to Fable 5.2 when your evals on Opus 5.5 at higher effort still fall short.

Key features

• Model ID claude-fable-5-2, the same string on Google Cloud, AWS and Foundry
• 1M token context window, 128K max output
• Batch API at 50% off
• Fast mode at $16 / $80 per million tokens
• Cache reads at $0.20 per million tokens

Best for

• Product teams that run their own agent loops or tools
• Teams that need batch jobs or fast mode on day one

Pricing $8 per million input tokens and $40 per million output tokens, 20% below Fable 5.1's $10 / $50. Fable 5.2 costs 2x Opus 5.5 per token, which is $4 / $20 (as of Nov 2026, Claude pricing).

Pros

• Every feature and price option is available here first
• Price cut versus Fable 5.1 on the same tasks

Cons

• 30-day retention for safety monitoring unless you qualify for Anthropic's enterprise safeguards option, so zero data retention setups need a check first
• High-risk cyber and most research biology requests are blocked and answered by older models instead

4. Google Cloud, best for companies that already bill on Google Cloud

Google Cloud offers Fable 5.2 through Model Garden in Vertex AI, which Google's docs now title Gemini Enterprise Agent Platform. Google's own Fable 5.2 model page lists the model as generally available, with a release date of November 9, 2026.

Pick this if your data and spend already sit in Google Cloud. You call the same claude-fable-5-2 ID, and you choose a US, EU or global endpoint.

Key features

• Launch stage: generally available
• Regions: us multi-region, eu multi-region and global
• 1,000,000 token context, 128,000 max output
• Can be opened in Agent Studio
• Fast mode with its own quota bucket

Best for

• Companies with committed spend on Google Cloud
• Teams that need EU data residency for model calls

Pricing Google's Fable 5.2 page links to its pricing page and does not list rates itself. It says fast mode costs 2x the standard rates (as of Nov 2026, Google Cloud docs).

Pros

• EU region on day one
• Billing and quotas sit inside your existing Google Cloud project

Cons

• Five setup steps: accept the Advanced AI Safety Addendum, click Enable on the model card, answer the use-case questions, accept the terms and turn on prompt-response sharing with Anthropic per location
• No rate is shown on the model page, so you have to look up the price separately

5. Claude Platform on AWS, best for teams that pay through their AWS bill

Claude Platform on AWS is Anthropic's own platform, sold and billed through your AWS account. Anthropic's docs say: "Unlike Amazon Bedrock, where AWS operates the inference stack, Anthropic operates Claude Platform on AWS." Its model table lists Claude Fable 5.2 as claude-fable-5-2.

AWS now offers Fable 5.2 in two ways: this platform, which Anthropic runs, and Amazon Bedrock, which AWS runs (entry 6). AWS's launch post says the model is on Claude Platform on AWS "in every Region where the service is offered."

Key features

• Same model IDs as the Claude API, with no anthropic. prefix or ARN
• Messages API, Agent Skills and code execution through your AWS account
• Python SDK client AnthropicAWS
• Endpoint in us-west-2 with SigV4 signing

Best for

• Companies with AWS committed spend who want Fable 5.2 now
• Teams that want the full Claude API feature set on an AWS invoice

Pricing Billed through AWS Marketplace and invoiced monthly in arrears on your AWS bill. The docs page we read does not list a Fable 5.2 rate, so check your Marketplace listing (as of Nov 2026, Anthropic docs).

Pros

• Anthropic runs it, so it follows the Claude API's features
• Code written for the Claude API works with a client swap

Cons

• A separate product from Bedrock, so Bedrock guardrails and IAM setups do not carry over
• Anthropic's docs do not list Mythos 5.2 on this platform

6. Amazon Bedrock, best for teams already building on Bedrock

Amazon Bedrock is AWS's managed model service, where AWS runs the inference stack. AWS's Fable 5.2 model card lists the model with a launch date of November 9, 2026, and AWS's launch post says it is "available today on Amazon Bedrock through US Geo CRIS and Global CRIS inference profiles."

There is no single-Region option. You call a cross-Region profile: us. keeps requests inside US Regions, global. can route them to any Region.

Key features

• Model ID anthropic.claude-fable-5-2, called as us.anthropic.claude-fable-5-2 or global.anthropic.claude-fable-5-2
• Messages, Converse and Invoke APIs on the bedrock-runtime endpoint
• Guardrails, Knowledge Bases, Agents, prompt caching and computer use, per the model card
• 1M token context, 128K max output

Regions

• us. profile: from us-east-1, us-east-2, us-west-1, us-west-2, ca-central-1 and ca-west-1, plus the two GovCloud Regions
• global. profile: from those 6 commercial Regions and 27 more, including Frankfurt, Ireland, London, Paris, Tokyo, Sydney, Mumbai and São Paulo
• No EU, Australia or Japan geo profile yet, so EU data residency is not available for Fable 5.2 on Bedrock

How to switch

1. Set your account or project data retention mode to aws_review with the Data Retention API. AWS requires it before the first call.
2. Give your IAM role bedrock:InvokeModel and bedrock:InvokeModelWithResponseStream.
3. Change modelId to global.anthropic.claude-fable-5-2, or to us.anthropic.claude-fable-5-2 to stay in the US.
4. Read the answer by block type, not by index. AWS notes a thinking block may come before the text block.

import boto3

client = boto3.client("bedrock-runtime", region_name="us-east-1")
response = client.converse(
    modelId="global.anthropic.claude-fable-5-2",
    messages=[{"role": "user", "content": [{"text": "Hello"}]}],
)

To try it without code, open the Bedrock console, go to the Playground and pick Fable 5.2.

Best for

• Teams that already run Bedrock Guardrails, Knowledge Bases or Agents
• AWS shops that want Fable 5.2 under their existing IAM policies

Pricing AWS's model card points to the Bedrock pricing page. That page did not show a Fable 5.2 rate when we read it on November 9, 2026, so check it before you budget. On the Claude API, Fable 5.2 is $8 / $40.

Pros

• Switching from another Claude model on Bedrock is a one-line modelId change
• Bedrock Guardrails, Knowledge Bases and Agents work with it, per AWS's model card

Cons

• Only us. and global. profiles, with no single-Region or EU option
• You must opt in to the aws_review data retention mode first
• Only the Standard service tier: Priority, Flex and Reserved are not supported yet

7. Vercel AI Gateway, best for apps on the Vercel AI SDK

Vercel AI Gateway is Vercel's API gateway that routes one key to many model providers. Its changelog from November 9, 2026 says: "Claude Fable 5.2 from Anthropic is now available on AI Gateway."

The useful part for production is fallbacks. Fable 5.2 can decline some requests through its safety classifier. Vercel lets you list backup models, such as Opus 5.5 and Sonnet 5.5, that answer when that happens.

Key features

• Model string anthropic/claude-fable-5.2 in the AI SDK
• A models fallback list under providerOptions.gateway
• Works with Chat Completions, Responses and Anthropic Messages APIs
• No platform fee on inference, including bring-your-own-key requests

Best for

• Next.js and AI SDK teams that already deploy on Vercel
• Apps that need a backup model when a request is declined

Pricing Vercel says AI Gateway "reflects provider pricing with no markup" (as of Nov 2026, Vercel changelog). Its model page lists $8 / $40 per million tokens.

Pros

• Fallback models in a few lines of config
• One key for Anthropic, OpenAI and Google models

Cons

• Fable 5.2 requires data retention, so teams with zero data retention turned on get an error
• Another vendor sits between you and Anthropic

8. OpenRouter, best for testing Fable 5.2 next to other models

OpenRouter is a model marketplace that serves many providers' models behind one API key. Its Fable 5.2 page was live when we read it on November 9, 2026, under the ID anthropic/claude-fable-5.2.

This is the quickest way to compare Fable 5.2 with models from other labs on your own prompts, without a new account at each one. The page lists Anthropic and 3 more providers.

Key features

• Model ID anthropic/claude-fable-5.2
• Several providers listed for the same model
• One key and one bill across labs
• OpenAI-compatible API

Best for

• Developers who run side-by-side model tests
• Small teams that want pay-as-you-go access with no enterprise contract

Pricing $8 per million input tokens and $40 per million output tokens with Anthropic as the provider (as of Nov 2026, OpenRouter).

Pros

• Easiest way to test across labs
• Same price as Anthropic direct on the Anthropic route

Cons

• The page shows no launch post or date, only the live listing
• Fast mode and batch pricing are not listed on the page, so use the Claude API if you need them

Which tools have not confirmed Fable 5.2 yet?

These tools and clouds are popular with Claude users. On November 9, 2026, none of them had a Fable 5.2 note on its own changelog or docs. Several usually post within a day or two of an Anthropic launch, so check back.

Gateways change the picture for some of these. Tools that accept an OpenRouter or Vercel AI Gateway key can likely reach anthropic/claude-fable-5.2 today. We do not list them until the tool itself says so.

How to choose the best way to run Claude Fable 5.2

1) Do you write code or build products?

If you write code, start in Claude Code. You switch with one command and see Fable 5.2 on your own repo within minutes. If you build a product, start on the Claude API, because it gets every feature and price option first.

2) Where does your company already pay for cloud?

If you have Google Cloud spend, use Google Cloud and pick the eu or us region you need. If you have AWS spend and already use Bedrock Guardrails or Knowledge Bases, use Amazon Bedrock. If you want the Claude API's features on an AWS bill, use Claude Platform on AWS.

3) Do you need fallbacks or many models?

If a declined request would break your app, use Vercel AI Gateway and list Opus 5.5 as the fallback. If you are still choosing a model, run the same 20 prompts through OpenRouter on Fable 5.2 and two other models, then compare cost per task, not price per token.

4) Is Fable 5.2 worth 2x Opus 5.5 for your work?

Run a 1-week test on your real tasks before you move everything. Anthropic's own docs say to start most workloads on Opus 5.5. On Anthropic's benchmark table, Fable 5.2 leads on FrontierSWE v2 (69.9% vs 62.3% for Opus 5.5), but Gemini 4 Carbon leads on AutomationBench. Count the total cost of a finished task.

Picking a model is one half of the work. The other half is being named when buyers ask ChatGPT, Claude or Perplexity which tools support it. Request a spot on AY Listicle to get your product into the "best of" lists that buyers and AI search read. We check every mention every month.

FAQ

Which tools support Claude Fable 5.2 right now?

Claude Code, the Claude apps, the Claude API, Google Cloud, Claude Platform on AWS, Amazon Bedrock, Vercel AI Gateway and OpenRouter support Claude Fable 5.2, according to each vendor's own changelog or docs on November 9, 2026. Cursor, GitHub Copilot and Microsoft's own docs had not confirmed it yet. What is the Claude Fable 5.2 model ID? The Claude Fable 5.2 model ID is claude-fable-5-2. The same ID works on the Claude API, Claude Platform on AWS, Google Cloud and Microsoft Foundry, per Anthropic's models overview. Amazon Bedrock uses anthropic.claude-fable-5-2, and Vercel AI Gateway and OpenRouter use anthropic/claude-fable-5.2. How much does Claude Fable 5.2 cost? Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens on the Claude API, 20% below Fable 5.1. Fast mode costs $16 / $80, and the Batch API is 50% off, per Anthropic's pricing page as of November 9, 2026. Pages that still show $10 / $50 describe Fable 5.1. Is Claude Fable 5.2 on Amazon Bedrock? Yes. AWS lists Claude Fable 5.2 on Amazon Bedrock as anthropic.claude-fable-5-2, launched November 9, 2026 (AWS model card). You call it through us.anthropic.claude-fable-5-2 or global.anthropic.claude-fable-5-2 after you set the aws_review data retention mode. There is no EU geo profile yet. Is Claude Fable 5.2 on Vertex AI? Yes. Google Cloud lists Claude Fable 5.2 as generally available in Model Garden, with the ID claude-fable-5-2 and us, eu and global endpoints. Google's docs now title the product Gemini Enterprise Agent Platform. Can I use Claude Mythos 5.2 in these tools? Claude Mythos 5.2 is available only to organizations verified through Anthropic's programs, such as the Cyber Verification Program and the Life Sciences Verification Program. Its ID is claude-mythos-5-2 and it runs on the Claude API, Google Cloud and Foundry for verified organizations, per Anthropic's release notes. AWS says Mythos 5.2 is also on Amazon Bedrock for authorized cyber and life sciences work (AWS). Claude Enterprise customers can reach it through Claude Security. Is Claude Fable 5.2 better than Opus 5.5 for coding? Claude Fable 5.2 scores higher than Opus 5.5 on Anthropic's own coding benchmarks, such as FrontierSWE v2 (69.9% vs 62.3%), at 2x the price per token. On the Vals Index, Vals AI's own leaderboard, Fable 5.2 ranks #1 of 36 at 71.31% (Vals AI). On the Artificial Analysis Intelligence Index v5.0.1, Claude Fable 5.2 (max) ranks #1 of 74 at 65, against 59 for Opus 5.5 (max) (checked November 9, 2026). When will Cursor and GitHub Copilot add Claude Fable 5.2? Neither Cursor nor GitHub Copilot had announced Claude Fable 5.2 on its own changelog by November 9, 2026, and both still list Fable 5 and Fable 5.1. We update this page when either posts a note.

Sources: Anthropic announcement, Amazon Bedrock, Claude Fable 5.2 model card, AWS, Introducing Claude Fable 5.2 on AWS, Artificial Analysis, Claude Fable 5.2, Claude models overview, Fable 5.2 model page, Claude pricing, Claude Platform release notes, Claude Code changelog, Claude Code model docs, Claude launch offers, Claude Platform on AWS docs, Google Cloud Fable 5.2 docs, Vercel changelog, OpenRouter, Vals AI. All read November 9, 2026. More on listicles and AI search on the AY Listicle blog, from AY Listicle.