TL;DR
Anthropic’s Claude Fable 5 is priced at $10 per million input tokens and $50 per million output tokens, double Claude Opus 4.8. Artificial Analysis data cited in the source material shows a 5.7% Intelligence Index gain over Opus 4.8, while the public productivity case rests on one unaudited Stripe customer story.
Anthropic’s Claude Fable 5, launched on June 9, 2026 and listed as generally available around July 1, is now the company’s most expensive active Claude model, priced at double Opus 4.8 while third-party benchmark data shows a 5.7% Intelligence Index gain.
Anthropic’s own pricing is the clearest part of the record: Claude Fable 5 is listed at $10 per million input tokens and $50 per million output tokens. That is exactly double the listed $5/$25 rate for Claude Opus 4.8 and more than three times the roughly $3/$15 rate for Sonnet 4.6 cited in the source material.
Artificial Analysis benchmark data puts Fable 5 at 64.9 on its Intelligence Index, compared with 61.4 for Opus 4.8. The same source lists a GDPval-AA knowledge-work Elo score of 1,932 for Fable 5 versus 1,890 for Opus 4.8.
The effective bill can be lower than the list rate. The source material cites a 90% prompt-caching discount, $5/$25 batch pricing, and a modeled $7.70 blended rate per million tokens at a 7:2:1 cache-hit/input/output mix. It also states that Fable 5 and Mythos 5 use the same underlying weights, with pricing kept the same and safeguards differing.
Budget Math Gets Harder
The price gap matters because AI budget owners pay premium rates for measurable gains in throughput, coding, analysis or harder tasks. On the aggregate figures supplied, Fable 5 costs 2.0x more than Opus 4.8 for a 5.7% Intelligence Index gain, and a full index run is estimated at $9,940 versus about $4,970.
That does not mean Fable 5 is weak. The source material says Anthropic and Artificial Analysis frame the launch as a number one result, and Fable 5 leads five of ten sub-benchmarks. For buyers, the practical question is whether their workload falls in the areas where larger task-level gains appear.
AI language model comparison
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How Fable Fits Claude Lineup
Fable 5 arrived after Opus 4.8 and Sonnet 4.6, with the source material placing its launch on June 9, 2026 and general availability around July 1, 2026. It is described as the most expensive model in Anthropic’s active Claude lineup, though older Opus 4/4.1 pricing at $15/$75 was higher.
The same source review says the pricing is supported by Anthropic’s launch post, platform pricing docs and product page, and repeated by outlets including Forbes, TechCrunch, Vellum, Finout and Artificial Analysis. The productivity record is thinner: the public evidence cited is a single Stripe customer example, not a controlled study.
“would otherwise have taken a whole team over two months by hand”
— Anthropic launch post, citing Stripe
best AI chatbot for business
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Productivity Claims Lack Audit Trail
The main unresolved issue is productivity proof. Anthropic’s launch post, according to the source material, credits Stripe with a codebase-wide migration in a 50-million-line Ruby codebase that took one day and would have taken a team more than two months by hand. That is a reported customer claim, not an audited human-baseline study.
It is also unclear how broadly the result applies across enterprise coding work, noncoding knowledge work, smaller teams or mixed-model deployments. The source review found zero controlled human-baseline studies and says stronger descriptions of the Stripe example as independently established proof were not supported.
AI model performance benchmark
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Buyers Wait for Workload Proof
The next test is outside the launch materials. Enterprise buyers, benchmarkers and analysts are likely to compare Fable 5 with Opus 4.8 on real workloads, with cost per completed task rather than model score alone. Any new audited customer data, Anthropic pricing update or expanded third-party benchmark set could change the value case.
Until then, procurement teams face a narrow choice: reserve Fable 5 for tasks where its lead is visible, keep Opus 4.8 or cheaper models for routine work, or run internal trials before moving 2026 AI budgets toward the premium tier.
AI token pricing tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Is Claude Fable 5 confirmed to cost twice Opus 4.8?
Yes. The source material says Fable 5 is priced at $10/$50 per million input/output tokens, while Opus 4.8 is listed at $5/$25.
Does the 5.7% gain mean workers are 5.7% more productive?
No. The 5.7% figure refers to an Artificial Analysis Intelligence Index gap, not a measured workplace productivity result.
What supports the Stripe productivity claim?
The source material says Anthropic cited a Stripe migration example, but it describes that evidence as a customer story, not an audited study.
Is Fable 5 the same as Mythos 5?
According to the source material, Fable 5 and Mythos 5 use the same underlying weights and have identical pricing, with differences in safeguards.
Source: Thorsten Meyer AI
