Some or all of the information on this page might not apply to Cloud de Confiance by S3NS. See Differences from Google Cloud for more details.

Anthropic's Claude Fable 5.2 on Google Cloud

Anthropic's Claude Fable 5.2 on Google Cloud is designed for technical and scientific work, including coding and finance. It can work on ambiguous problems over long periods, verify its own work, and orchestrate subagents.

Retirement date: No sooner than November 9, 2027.

Try in Agent Studio View model card in Model Garden

Technical specifications
Model ID	`claude-fable-5-2`
Launch stage	GA
Supported inputs & outputs	Inputs: Text, Image, PDF Outputs: Text
Token limits	Maximum input tokens: 1,000,000 Maximum output tokens: 128,000
Capabilities	Supported Fast mode Computer use Web search Batch predictions Prompt caching Function calling Count tokens Memory tool Not supported
Usage types	Supported Shared Model Lineage Quota Provisioned Throughput Not supported
	Images	Limitation and specifications: See Vision in Anthropic's documentation
	Documents	Limitation and specifications: See PDF support in Anthropic's documentation
Versions	`claude-fable-5-2` Launch stage: Generally available Release date: November 9, 2026 Retirement date not sooner than: November 9, 2027
Supported regions
	Model availability (Includes fixed quota & Provisioned Throughput)	United States `us (US multi-region)` Europe `eu (Europe multi-region)` Global `global (global endpoint)`
	ML processing	United States `Multi-region` Europe `Multi-region`
Quota limits	us (US multi-region): QPM: Granted upon model enablement (Shared lineage quotas) TPM: Granted upon model enablement (Shared lineage quotas) Context length: 1,000,000 eu (Europe multi-region): QPM: Granted upon model enablement (Shared lineage quotas) TPM: Granted upon model enablement (Shared lineage quotas) Context length: 1,000,000 global (global endpoint): QPM: Granted upon model enablement (Shared lineage quotas) TPM: Granted upon model enablement (Shared lineage quotas) Context length: 1,000,000
Pricing	See Pricing.

Fast mode

Anthropic's Claude Fable 5.2 on Google Cloud supports fast mode to accelerate output token generation for latency-sensitive, interactive workloads. Set "speed": "fast" in online prediction requests to claude-fable-5-2:

Pricing: Fast mode is priced at 2x the standard-speed token rates across all token types (input, output, 5-minute and 1-hour cache write, and cache hit tokens). Web search tool calls are billed at the standard rate. See the Agent Platform pricing page.
Quotas: On global and multi-region endpoints, fast mode requests draw from a dedicated anthropic-claude-fable-fast shared lineage quota bucket separate from standard anthropic-claude-fable quota.
No Provisioned Throughput: Fast mode doesn't support Provisioned Throughput. Fast mode requests always use pay-as-you-go, and fast mode requests sent with the X-Vertex-AI-LLM-Request-Type: dedicated header are throttled with an HTTP 429 (RESOURCE_EXHAUSTED) error.

For code samples, monitoring, and limitations, see Fast mode with Anthropic Claude models.

Enable Anthropic's Claude Fable 5.2 on Google Cloud in your project

Follow these steps to enable Anthropic's Claude Fable 5.2 on Google Cloud in your project:

Consent to the Advanced AI Safety Addendum in Model Garden (required once per project). See Abuse Monitoring to learn more.
For online terms customers, click Enable on the Model Garden model card to:
- Answer relevant company and use case questions in Model Garden.
- Accept Anthropic and Marketplace terms of service in Cloud Marketplace.
For Private Offer customers:
- Accept the Private Offer in Cloud Marketplace (required once per Billing Account).
- Enable the model location using the enableModel API.
As required by Section F of Anthropic's Service Specific Terms, enable prompt-response sharing with Anthropic for each model location. See Request-Response Logging and Sharing to learn more.