🔍 Read the full analysis: A 5.7% AI Intelligence Increase With A Twofold Price Tag on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Anthropic lists Claude Fable 5 at $10 per million input tokens and $50 per million output tokens, double the price of Opus 4.8. Artificial Analysis reports a 5.7% higher Intelligence Index score, but the source material identifies no controlled human-baseline productivity study and cites one unaudited customer account.
Anthropic’s Claude Fable 5 costs twice as much as Claude Opus 4.8, while scoring 5.7% higher on Artificial Analysis’s Intelligence Index, according to a July 2026 review of model pricing and performance. The review says the public case for increased workplace productivity rests on one unaudited customer account, not a controlled study, leaving buyers with clearer evidence about cost than about business returns.
Anthropic lists Fable 5 at $10 per million input tokens and $50 per million output tokens. Opus 4.8 is priced at $5 and $25 per million tokens, respectively. Sonnet 4.6 is listed at approximately $3 and $15, making Fable 5 about 3.3 times as expensive on those rates. The source review says Anthropic’s launch post, pricing documentation and product page support Fable 5’s list price.
Artificial Analysis reports an Intelligence Index score of 64.9 for Fable 5, compared with 61.4 for Opus 4.8, a 5.7% difference as calculated in the review. On GDPval-AA, a knowledge-work benchmark reported as an Elo rating, Fable 5 scores 1,932 against Opus 4.8’s 1,890. The review notes that an Elo gap should not be read as an ordinary percentage difference: it says the 42-point lead corresponds to roughly a 56% head-to-head win rate under the rating interpretation it cites.
The comparison is not uniformly unfavorable to the more expensive model. The review says Fable 5 set records in five of 10 sub-benchmarks, with larger gains on some tasks than its overall index result suggests. It also estimates that a full Intelligence Index run costs about $9,940 for Fable 5 and $4,970 for Opus 4.8. Those benchmark and cost figures describe measured tests, not a forecast of what a particular organization will save or produce.
The Cost of a Small Average Gain
For organizations choosing between models, the figures sharpen a practical question: does the more expensive system perform better on the work employees actually do? A higher aggregate benchmark score may matter for demanding tasks, but it does not by itself show that a company will complete work faster, reduce staffing costs or improve output quality enough to cover the price premium.
The reported index gap is an average across a benchmark, and the review says results vary across its sub-benchmarks. That makes workload fit central to the purchasing decision. A team whose tasks resemble areas where Fable 5 performs especially well may judge its price differently from a team handling work where the models are closer. The available material does not establish which outcome applies to a given buyer.
The distinction also matters for budgeting. Token prices can be checked directly, while productivity claims require a defined baseline, a measurement period and a method for separating the model’s contribution from other changes. Without those details, a company riskily treats a customer story as if it were a general, independently verified return-on-investment finding. The review’s central conclusion is narrower: the price is documented; broad productivity gains are not demonstrated in the cited evidence.
AI language model pricing comparison
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What the Benchmarks Measure
Fable 5 is described in the source material as Anthropic’s most expensive model in its active lineup. The listed rates are not the only available rates: the review reports a 90% prompt-caching discount, cache reads at $1 per million tokens, and batch pricing of $5 per million input tokens and $25 per million output tokens. It calculates a blended rate of $7.70 per million tokens using a particular 7:2:1 mix of cache-hit, input and output tokens. That is an illustrative mix, not a universal customer rate.
The review also says US-only inference is available at a 1.1-times price multiplier, or $11 per million input tokens and $55 per million output tokens. It states that Fable 5 and Mythos 5 use the same underlying weights and have identical prices, with safeguards distinguishing the products. These details affect how buyers may compare effective costs, but they do not change the review’s headline comparison of standard list prices.
On productivity, the source identifies a launch-post account attributed to Stripe: Anthropic said Fable 5 completed a migration across a 50-million-line Ruby codebase in one day, a task that otherwise would have taken a team more than two months by hand. The review confirms that this statement appeared as a claim in Anthropic’s launch material. It says the account has no disclosed methodology, audit or independent replication in the material it examined. That makes it a customer example, not a controlled measurement of typical enterprise results.
“In a 50-million-line Ruby codebase, Fable 5 performed a codebase-wide migration in a day that would otherwise have taken a whole team over two months by hand.”
— Anthropic, in its launch post, describing a Stripe customer example
enterprise AI model performance benchmarks
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Missing Workplace Baseline
The reviewed material reports no controlled human-baseline productivity study for Fable 5 and one enterprise customer example that the review calls unaudited. It does not provide independent replication, a detailed measurement method or evidence that the reported result generalizes to other companies, codebases or kinds of work. The account shows what Anthropic said happened in one project; it does not establish an expected productivity gain for a typical customer.
It is also unclear how the benchmark advantage translates into the cost of real deployments. Actual bills depend on usage, caching, task length and the balance of input and output tokens. The $7.70 blended estimate relies on a stated token mix, while the approximately $9,940 and $4,970 figures refer to the cost of full benchmark runs in the review. Neither figure is a general estimate of monthly enterprise spending.
The source material does not provide a complete task-by-task comparison, the underlying test methodology for every reported result, or a verified comparison of Fable 5’s return on investment across customers. Its mention of a roughly twofold input-price comparison with GPT-5.5 is based on a single secondary source and is explicitly described as directional. The clearest supported comparison remains Fable 5 against Opus 4.8 on the stated prices and benchmarks.
As an affiliate, we earn on qualifying purchases.
Evidence Buyers Should Seek
For now, buyers considering Fable 5 can compare the published rates with their own expected token use and test the model on representative tasks. A useful evaluation would set out a baseline using the current process or a lower-cost model, measure time and output quality over a defined period, and account for review, correction and integration work. Those results would help establish whether gains on the relevant workload justify the added spend.
The source review does not identify a forthcoming independent productivity study or a confirmed date for one. Further customer disclosures, controlled evaluations or independently audited results could change how confidently organizations assess the return. Until such evidence is available, the benchmark results support a claim of stronger performance on specified tests, while broad claims about workplace productivity remain unverified in the material cited.
AI productivity benchmarking software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How much does Claude Fable 5 cost?
Anthropic’s listed rates are $10 per million input tokens and $50 per million output tokens. The source review also reports discounts for prompt caching and batch processing, so actual costs can differ by usage.
How does Fable 5 compare with Opus 4.8?
Fable 5 costs twice as much at the listed rates. Artificial Analysis reports an Intelligence Index score of 64.9 versus 61.4, a 5.7% difference as calculated in the review. The review also reports a 1,932-to-1,890 score on GDPval-AA.
Has Fable 5’s productivity advantage been independently proven?
Not in the evidence described by the source review. It identifies no controlled human-baseline productivity study and cites one Stripe customer account from Anthropic’s launch post. The review says that account has not been independently audited or replicated in the material it assessed.
Could Fable 5 still be worth the higher price?
That depends on the workload and actual usage. The review says Fable 5 performed especially well in five of 10 sub-benchmarks, but the results do not establish that every organization will get enough productivity or quality improvement to offset the higher price.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
