Reviewed by Jonathan West · Updated Nov 9, 2026

Claude Fable 5.2 Pricing

Complete per-token rate card, cache discounts, plan inclusions, and internal cost modeling across production workloads.

Reviewed by Jonathan West · Updated Nov 9, 2026

Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, with prompt cache reads priced at $0.20 and Batch Application Programming Interface (API) requests at $4 and $20. Released by Anthropic on November 9, 2026, these figures represent a 20% reduction across all standard token rates compared to Claude Fable 5.1.

A request with 100,000 input tokens and 5,000 output tokens costs exactly $1.00. The underlying architecture provides a 1M-token context window billed without a long-context penalty, while optional execution modes such as fast mode and prompt caching alter final charges.

Evaluating Claude Fable 5.2 pricing requires balancing lower base rates against adaptive thinking token usage and deployment configurations. Organizations evaluating migrations from Claude Fable 5.1 or Claude Opus 5.5 must calculate cost per task rather than looking solely at baseline rate cards.


Claude Fable 5.2 Pricing per Million Tokens

Anthropic publishes dedicated rates for prompt caching, asynchronous batch processing, and fast mode inference across its developer platforms. Standard prompt caching lets developers write a prompt to cache for five minutes at $10 per million tokens. A one-hour write costs $16 per million tokens. Cached prompts read back at $0.20 per million tokens, which equals 0.025 times the standard input rate.

The Batch API cuts standard input and output prices by 50%, charging $4 per million input tokens and $20 per million output tokens for non-urgent processing. Fast mode doubles the per-token rate to $16 per million input tokens and $80 per million output tokens in exchange for lower latency.

Rate TypeInput Tokens (per MTok)Output Tokens (per MTok)Details
Standard API$8.00$40.00Base interactive inference across all supported clouds
Prompt Cache Read$0.20N/A0.025 times base input rate on cached prefixes
5-Minute Cache Write$10.00N/A1.25 times the base input rate; cache valid for 5 minutes
1-Hour Cache Write$16.00N/A2 times the base input rate; cache valid for 1 hour
Batch API$4.00$20.0050% discount for asynchronous processing
Fast Mode$16.00$80.00Up to 2.5 times higher output tokens per second

What Changes the Price

Several API parameters directly modify the effective cost per million tokens (MTok) when querying Claude Fable 5.2.

Batch processing cannot be combined with fast mode.

Fast mode accelerates generation up to 2.5 times higher output tokens per second by passing speed: "fast" in API calls. Fast mode is supported on the Claude API, Claude Models on Google Cloud, and Claude in Microsoft Foundry (as a separate model claude-fable-5-2-fast).

Geographic routing rules introduce an additional multiplier. Selecting United States (US) only inference via inference_geo applies a 1.1 times multiplier across all base, cache, and fast mode token rates, whereas global routing retains standard list rates.

Anthropic applies no long-context pricing penalty on Claude Fable 5.2. A 900k-token request within the 1M-token context window is billed at the standard $8 per million input token rate.

Developers calculating custom request combinations can review the exact API parameters on the official Claude API pricing documentation.


How Fable 5.2 Compares With Fable 5.1, Opus 5.5 and Sonnet 5.5 on Price

Claude Fable 5.2 is positioned directly between Anthropic's daily driver models and its predecessor generations on overall price.

Against its predecessor, Claude Fable 5.2 reduces standard pricing by 20%. Claude Fable 5.1 pricing stands at $10 per million input tokens and $50 per million output tokens, with cache writes at $12.50 (5-minute) and $20 (1-hour), and cache reads at $0.25. The price drop applies across every billing dimension, including batch and caching.

When compared to Anthropic's recommended default model, Claude Opus 5.5 pricing is exactly half the cost of Claude Fable 5.2. Claude Opus 5.5 charges $4 per million input tokens and $20 per million output tokens, with batch rates of $2 and $10. Anthropic advises engineering teams to begin prototyping on Claude Opus 5.5 and escalate to Claude Fable 5.2 only when evaluations at high effort fail to meet accuracy targets.

Further down the catalog, Claude Sonnet 5.5 costs $2 per million input tokens and $10 per million output tokens. Claude Haiku 5.5 starts from $0.10 input and $0.50 output per million tokens.

Claude Mythos 5.2 shares identical rates with Claude Fable 5.2, costing $8 per million input tokens and $40 per million output tokens. Access to Claude Mythos 5.2 is restricted to verified organizations through programs such as the Cyber Verification Program and Life Sciences Verification Program, as detailed in our guide on Claude Fable 5.2 vs Mythos 5.2.


What Claude Plans Cost and Whether Fable Is Included

Claude consumer and workplace subscription tiers carry different permanent entitlements and promotional allowances for Claude Fable 5.2.

On Anthropic's permanent plan catalog, the Free tier costs $0 per month and does not include Claude Fable 5.2 access. Free tier users are limited to Claude Sonnet and Claude Haiku. The Pro plan costs $17 per month with an annual billing discount or $20 billed monthly. Its permanent entitlement requires purchasing usage credits to run Claude Fable models, while Claude Opus is included.

The Claude Max plan at $100 per month and Claude Ultra plan at $200 per month include Claude Fable up to 50% of weekly usage limits. Team plans priced on standard tiers require usage credits for Claude Fable, whereas premium Team seats include Claude Fable up to 50% of weekly limits. Claude Enterprise organizations receive Claude Fable access directly.

To mark the release, Anthropic announced a dated launch promotion running from November 9, 2026 through November 16, 2026 at 11:59 pm Pacific Time (PT). During this window, Pro plans and standard seats on Team plans with billing in place prior to November 9, 2026 at 10:00 am PT receive all Claude Fable models included up to 50% of weekly usage limits, plus one usage reset that expires on November 16, 2026 at 11:59 pm PT. After this promotional window closes, or after the 50% limit is consumed, accounts revert to paying via usage credits.

Subscribers on Max, Ultra, premium Team seats, and premium seats on legacy seat-based Enterprise plans receive one to three usage resets based on subscription length, which expire on March 8, 2027 at 11:59 pm PT. For an analysis of free access avenues, review our guide covering Is Claude Fable 5.2 Free?.


Our Modeled Monthly Cost for Three Workloads

In our internal cost modeling, we calculated monthly API expenses across three assumed workloads using Anthropic's published rates to evaluate the financial difference between Claude Fable 5.2, Claude Fable 5.1, and Claude Opus 5.5.

At Layer3Labs, we modeled these estimates assuming identical prompt and completion counts, giving a baseline comparison of rate differences across varying volumes and processing modes.

The first workload models a high-throughput support-chat API handling 2,000,000 requests per month. Each request averages 3,000 input tokens and 500 output tokens, totaling 6,000,000,000 input tokens and 1,000,000,000 output tokens. Under standard interactive rates, Claude Fable 5.2 costs $88,000.00 per month, compared to $110,000.00 on Claude Fable 5.1 and $44,000.00 on Claude Opus 5.5.

The second workload evaluates a specialized engineering pipeline running 200 coding agent hours per month. Each agent session uses 1,500,000 input tokens and generates 150,000 output tokens, totaling 300,000,000 input tokens and 30,000,000 output tokens. Claude Fable 5.2 costs $3,600.00 per month for this volume, whereas Claude Fable 5.1 costs $4,500.00 and Claude Opus 5.5 costs $1,800.00.

The third workload evaluates an asynchronous batch document job processing 50,000 files monthly. Each file averages 20,000 input tokens and generates 1,500 output tokens, yielding 1,000,000,000 input tokens and 75,000,000 output tokens. Using the 50% Batch API discount, Claude Fable 5.2 totals $5,500.00 per month, compared to $6,875.00 on Claude Fable 5.1 and $2,750.00 on Claude Opus 5.5.

Workload ScenarioAssumed Monthly VolumeClaude Fable 5.2 CostClaude Fable 5.1 CostClaude Opus 5.5 Cost
Support-Chat API (Interactive)2,000,000 requests (3,000 in / 500 out)$88,000.00$110,000.00$44,000.00
Coding Agent Hours (Interactive)200 agent hours (1,500,000 in / 150,000 out)$3,600.00$4,500.00$1,800.00
Batch Document Job (Batch API)50,000 files (20k in / 1.5k out)$5,500.00$6,875.00$2,750.00

Because Claude Fable 5.2 token rates are exactly 2.0 times higher than Claude Opus 5.5, a break-even threshold exists at a token consumption ratio of 0.5. Claude Fable 5.2 costs more than Claude Opus 5.5 whenever it consumes more than 50% of the token volume that Claude Opus 5.5 requires to complete the same task.

Engineering teams sizing their deployments can compare figures against our AI model pricing table or simulate customized input and output mixes directly using our AI model cost calculator.

Budget Claude Fable 5.2 at $8 and $40 per million tokens only for tasks where measured performance justifies the expense, and test your planned request volumes in the cost calculator before migrating production workloads to Claude Fable 5.2.

Frequently Asked Questions

  • Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens for standard API requests. Prompt cache reads cost $0.20 per million tokens, while the Batch API charges $4 per million input tokens and $20 per million output tokens.
  • Claude Mythos 5.2 shares the exact same pricing as Claude Fable 5.2, billing at $8 per million input tokens and $40 per million output tokens. Access to Claude Mythos 5.2 requires specialized organizational verification rather than a different commercial rate card.
  • Claude Fable 5.2 carries no long-context pricing surcharges across its entire 1M-token context window. A 900k-token request is billed at the same per-token rate as a 9k-token request.
  • The Batch API provides an automatic 50% discount on both input and output rates, pricing Claude Fable 5.2 at $4 per million input tokens and $20 per million output tokens. Batched jobs are processed asynchronously and cannot be paired with fast mode.
  • Anthropic does not offer a free tier for Claude Fable 5.2. Free accounts on Claude apps are limited to Claude Sonnet and Claude Haiku models, while Claude Fable access requires an eligible paid plan, promotional allowance, or API credits.

Review Your Model Deployment Costs

Layer3Labs helps engineering teams analyze token spend and optimize Claude Fable 5.2 pricing for production systems. Schedule a technical review to evaluate whether your workloads warrant the move.

Book a Consultation