TL;DR
Anthropic’s Claude Fable 5 is the company’s most expensive active public model, priced at $10 per million input tokens and $50 per million output tokens. Independent benchmark data cited in the source material shows a 5.7% Intelligence Index gain over Opus 4.8, which costs half as much, while direct productivity evidence remains limited to one unaudited customer story.
Anthropic’s Claude Fable 5 is now positioned as the company’s most expensive active public model, charging $10 per million input tokens and $50 per million output tokens, while independent benchmark data cited in the source material shows only a 5.7% Intelligence Index gain over Claude Opus 4.8, which costs half as much.
The confirmed pricing comes from Anthropic materials cited in the source: Fable 5 lists at $10/$50 per million input and output tokens. That is 2.0 times Opus 4.8, listed at $5/$25, and about 3.3 times Sonnet 4.6, listed near $3/$15. The same source says Fable 5 and Mythos 5 share the same underlying weights and price, with differences in safeguards.
The benchmark comparison is narrower. According to Artificial Analysis data cited in the source, Fable 5 scored 64.9 on the Intelligence Index, compared with 61.4 for Opus 4.8, a 5.7% increase. On GDPval-AA, described as a knowledge-work Elo benchmark, Fable 5 scored 1,932 versus 1,890 for Opus 4.8. The source also reports that one full Intelligence Index run costs about $9,940 for Fable 5, compared with about $4,970 for Opus 4.8.
The source separates those measured results from productivity claims. It says there are no controlled human-baseline productivity studies in the public record for Fable 5. The surviving enterprise proof point is one unaudited customer account, attributed to Anthropic’s launch material and credited to Stripe.
Budget Pressure Meets Thin Proof
The gap matters because AI budget decisions often turn on whether a top-priced model produces enough extra output to justify its rate. Here, the confirmed cost increase is large: double Opus 4.8 for input and output tokens. The confirmed average benchmark gain cited in the source is smaller: 5.7% on the Intelligence Index.
That does not mean Fable 5 is weak. The source describes it as the highest-scoring public Anthropic model on aggregators and says it set records in five of 10 sub-benchmarks. But the business question is narrower: whether those gains match a buyer’s workload. A composite score can hide larger gains on some tasks and flatter results on others.
For enterprises, the main risk is overstating productivity. The source says the pricing is documented through primary and secondary references, while the productivity case rests on a single customer story. That leaves finance, engineering and procurement teams with solid cost data but limited public evidence on human work saved.
Evals for AI Engineers: Systematically Measuring and Improving AI Applications
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How Fable Compares Inside Claude
Fable 5 entered the public Claude lineup after a June 9, 2026 launch and was described as generally available around July 1, 2026. In that lineup, it sits above Opus 4.8 and Sonnet 4.6 on price. Older Opus 4 and 4.1 models were listed higher at $15/$75, but the source says they are retired.
The effective price can be lower than the headline rate when caching or batch pricing applies. The source lists a 90% prompt-caching discount, cache reads at $1 per million tokens, and batch pricing at $5/$25. Using a 7:2:1 mix of cache hits, fresh input and output, the cited blended rate is $7.70 per million tokens.
Some comparisons remain less firm. The source says Fable 5’s input price is roughly twice GPT-5.5’s, but describes that comparison as based on one secondary source. That figure should be treated as directional unless primary pricing confirms it.
AMT Star Trek Model Kit (U.S.S. Enterprise Refit)
Plastic Model Kit, Assembly required
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Productivity Evidence Remains Limited
It is not yet clear whether Fable 5’s benchmark lead translates into broad workplace productivity gains. The source says there are zero controlled human-baseline studies available publicly and only one unaudited enterprise proof point. That means the strongest productivity framing remains a claim made in launch material, not a replicated result.
It is also unclear which customers would see enough benefit to justify the premium. The source says Fable 5 performs better on some sub-benchmarks, but the average gain over Opus 4.8 is modest relative to the price increase. Buyers still need workload-specific testing before treating the model as a default upgrade.
AI model pricing and performance analysis
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Buyers Need Workload Tests
The next step for customers is likely side-by-side evaluation: run Fable 5 against Opus 4.8 and Sonnet 4.6 on real tasks, measure completion quality, latency, token use and review time, then compare those results with the confirmed rate card. Public evidence may also change if Anthropic releases audited case studies or if independent evaluators publish more task-level results.
Until then, the confirmed picture is mixed: Fable 5 is Anthropic’s top public model by cited aggregate scores, but the public business case for paying the premium remains unproven beyond limited evidence.
large language model input output tokens
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?
Claude Fable 5 is listed at $10 per million input tokens and $50 per million output tokens. The source also cites cache and batch discounts that can lower the effective rate for some workloads.
How much better is Fable 5 than Opus 4.8?
According to Artificial Analysis data cited in the source, Fable 5 scored 64.9 on the Intelligence Index versus 61.4 for Opus 4.8, a 5.7% gain.
Is there proof that Fable 5 makes workers more productive?
The source says no controlled human-baseline productivity studies are publicly available. The main public example is one unaudited Stripe story cited by Anthropic.
Is Fable 5 always the best Claude model to buy?
Not based on public evidence alone. Fable 5 leads on cited aggregate benchmarks, but Opus 4.8 costs half as much. The better choice depends on whether a specific workload benefits from Fable 5’s stronger sub-benchmark areas.
Are Fable 5 and Mythos 5 different models?
The source says Fable 5 and Mythos 5 use the same underlying weights and have the same price, with differences tied to safeguards.
Source: Thorsten Meyer AI
