🔍 Read the full analysis: How The AI Price Doubles For A 5.7% Intelligence Gain 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 prices Claude Fable 5 at $10 per million input tokens and $50 per million output tokens, twice the rates for Opus 4.8. Artificial Analysis reports a 5.7% lead on its Intelligence Index, but the available human-baseline productivity evidence described in the source is one unaudited customer anecdote, not a controlled study.
Anthropic’s Claude Fable 5 costs twice as much per token as Opus 4.8, while scoring 5.7% higher on Artificial Analysis’s Intelligence Index, according to a July 2026 analysis of model pricing and benchmark data. The same analysis says the public case for Fable 5 improving enterprise productivity rests on one unaudited customer account, rather than a controlled study measuring work against a human or no-AI baseline.
Anthropic’s listed price for Fable 5 is $10 per million input tokens and $50 per million output tokens. Opus 4.8 is listed at $5 and $25, respectively, while Sonnet 4.6 is approximately $3 and $15, according to the source analysis, which says the prices are corroborated by Anthropic’s own materials and third-party reporting. Fable 5 is the most expensive model in Anthropic’s active lineup, based on those rates.
On Artificial Analysis’s Intelligence Index, Fable 5 scored 64.9 compared with Opus 4.8’s 61.4, a 5.7% difference as reported by the analysis. On GDPval-AA, a benchmark for knowledge-work tasks, Fable 5 scored 1,932 Elo points against Opus 4.8’s 1,890, a 42-point lead that the source reports as 2.2%. The analysis cautions that this percentage alone may not convey the meaning of an Elo gap; it says the benchmark’s logarithmic rating corresponds to an estimated 56% head-to-head win rate.
Artificial Analysis also found Fable 5 set records in five of 10 sub-benchmarks, according to the source. That may matter to buyers whose tasks align with those tests, but it does not establish that every business will see a similar advantage. The analysis estimates a full Intelligence Index run at about $9,940 for Fable 5 and $4,970 for Opus 4.8, reflecting the models’ two-to-one price difference.
The Cost of a Benchmark Lead
The pricing gap gives companies a concrete cost to weigh against benchmark performance. For a buyer choosing between Fable 5 and Opus 4.8, the reported results show a higher score at twice the listed token price; whether that premium is worthwhile depends on the buyer’s workload, quality needs and actual usage costs. Benchmark scores can help compare models, but they do not directly tell an organization how much time or money a deployment will save.
That distinction is central to the source analysis. It says Fable 5 is the highest-scoring public model Anthropic has released on the cited aggregate index, while arguing that claims about workplace productivity remain much less documented than the price. A company budgeting for the model should not treat a benchmark lead or a launch testimonial as a guaranteed productivity return. The available evidence, as described in the source, does not provide an audited, controlled estimate of enterprise-wide gains.
The business case may still be strong for some teams. The analysis notes that average benchmark results can obscure better performance on particular tasks: Fable 5’s records across five sub-benchmarks could make it more useful for work resembling those evaluations. But organizations would need to test their own workloads and measure costs and output quality to determine whether those task-specific differences justify the premium.
As an affiliate, we earn on qualifying purchases.
Prices, Caching and Launch Claims
The source analysis says Fable 5 launched on June 9, 2026, with general availability reported around July 1. Anthropic’s launch materials and pricing pages are cited as the primary basis for its rates. The analysis also says Fable 5 and Mythos 5 use the same underlying weights and are priced identically, with their difference described as safeguards rather than model weights.
List prices are not the only possible cost. The source 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 specified mix: seven cache-hit units, two input units and one output unit. That is an illustrative usage ratio, not a universal rate. The source also reports a 1.1-times multiplier for US-only inference.
The productivity example highlighted in Anthropic’s launch post concerns a 50-million-line Ruby codebase. The post attributes to Stripe the claim that Fable 5 completed a codebase-wide migration in a day, a job that otherwise would have taken a team more than two months by hand. The source analysis says this statement was verified as a claim made in the launch post, but not as an independently measured or replicated result.
As an affiliate, we earn on qualifying purchases.
What Productivity Data Is Missing
The source analysis reports no controlled Fable 5 knowledge-work productivity study against a human or no-AI baseline and identifies one enterprise human-baseline example: the Stripe account in Anthropic’s launch post. It describes that account as unaudited and says no methodology or independent replication was available in the material reviewed. The anecdote is evidence that a customer claim was made; it does not establish a general productivity rate.
The source does not provide enough information to determine how the migration was scoped, how much human review or preparation was involved, what counted as completion, or whether the reported timeline can be reproduced elsewhere. It also does not establish that Fable 5’s benchmark advantage translates into a comparable advantage on a reader’s own work. The benchmark results are specific measurements, while business outcomes depend on task mix, implementation, oversight, token use and the cost of correcting errors.
Other comparisons also need care. The analysis says its comparison with GPT-5.5 relies on a single secondary source and should be treated as directional. Its blended Fable 5 rate depends on an assumed token mix, so actual bills could differ substantially. The source does not give an independently audited estimate of total deployment costs or savings for customers.
As an affiliate, we earn on qualifying purchases.
The next step for organizations considering Fable 5 is to evaluate it on representative work rather than infer expected returns from a launch anecdote or aggregate benchmark alone. A useful comparison would track task completion, quality, human review time and total token costs for Fable 5 and a lower-priced alternative, using the same tasks and evaluation criteria. The source provides no schedule for an independent productivity study or additional customer evidence.
Further reporting could clarify whether Anthropic or its customers will publish methods and results that allow the Ruby migration claim to be assessed, and whether more enterprise examples will include measured baselines. Until then, the public record described in the July analysis supports a clear conclusion on listed prices and benchmark scores, but not a general estimate of productivity gains. Buyers should treat the return on the price premium as an open question to test in their own operations.
As an affiliate, we earn on qualifying purchases.
Key Questions
How much does Claude Fable 5 cost?
The source analysis reports Anthropic list prices of $10 per million input tokens and $50 per million output tokens. It also reports lower cache and batch rates, which apply under their respective terms.
How does Fable 5 compare with Opus 4.8 on benchmarks?
Artificial Analysis’s Intelligence Index score was 64.9 for Fable 5 and 61.4 for Opus 4.8, a reported 5.7% difference. On GDPval-AA, the scores were 1,932 and 1,890 Elo, respectively. These are benchmark results, not direct measurements of productivity at a particular company.
Is there proof that Fable 5 makes workers more productive?
The source analysis identifies one public human-baseline example: a Stripe migration account repeated in Anthropic’s launch post. It says the account is unaudited and not independently replicated, and reports no controlled Fable 5 productivity study against a human or no-AI baseline.
Does the benchmark lead justify paying twice as much?
There is no universal answer in the available evidence. The value depends on whether a company’s tasks benefit from Fable 5’s strengths and whether those gains outweigh its higher usage costs. The source analysis says Fable 5 led on five of 10 cited sub-benchmarks, but buyers would need to test their own work to assess the premium.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
