🔍 Read the full analysis: What You’ll Pay Per Task At Each Claude Fable 5.2 Effort Level on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
Artificial Analysis testing puts Claude Fable 5.2’s average API cost at $3.10 per task on Low effort and $26.54 on Max, while its Intelligence Index score rises from 54 to 65. High scores 62 for $9.08 per task; Max costs nearly three times as much for three additional points. The figures are averages from the index’s evaluations, not a guaranteed price for every real-world task.
Claude Fable 5.2’s estimated cost per task ranges from $3.10 at Low effort to $26.54 at Max, according to Artificial Analysis testing published November 10. The benchmark assigns the model scores of 54 and 65 at those respective settings, giving businesses a concrete estimate of the price of asking the new Anthropic model to spend more effort on a task.
Artificial Analysis tested five settings—Low, Medium, High, Xhigh and Max—on its Intelligence Index, which includes nine evaluations covering agentic coding, business documents, science and knowledge. The reported “cost per task” is the average API cost for one task in that test. The results are $3.10 at Low for a score of 54, $3.80 at Medium for 56, $9.08 at High for 62, $14.83 at Xhigh for 62, and $26.54 at Max for 65. The index ranks those settings 15th, ninth, third, second and first, respectively, among 74 model entries.
The estimates use Anthropic’s listed API rates of $8 per million input tokens and $40 per million output tokens for Fable 5.2. The list price is the same at every effort setting; the bill changes because the model uses different amounts of tokens. Across the full index, Artificial Analysis reports output-token use rising from 34.9 million at Low to 275 million at Max. Medium is the default setting in Claude apps and Claude Code, according to the source report.
The cost increase is not matched by equal score gains at every step. Moving from Medium to High adds six index points for $5.28 more per task. Xhigh costs about 63% more than High but has the same displayed score of 62. Max adds three points over High while costing nearly three times as much, and averages 46.4 minutes per task, compared with 17.9 minutes at High and 6.5 at Low. Artificial Analysis reports that on its FrontierCode 2.0 coding test, Fable scored 49.1% at High and 48.8% at Max.
What Fable 5.2 costs at each effort level
Artificial Analysis estimates average task costs rise with effort, while benchmark scores improve unevenly. These are index averages, not guaranteed prices for real requests.
Cost and score by setting
Higher effort uses more tokens; the listed token prices are the same across settings.
| Effort | Cost per task | Score |
|---|---|---|
| Low | $3.10 | 54 |
| Medium | $3.80 | 56 |
| High | $9.08 | 62 |
| Xhigh | $14.83 | 62 |
| Max | $26.54 | 65 |
The trade-off at the top
The extra spend and time do not produce equal gains at every step.
A lower-cost benchmark result
High scores 62 for $9.08 per task and averages 17.9 minutes.
A small score gain
Max scores 65 for nearly three times High’s cost and averages 46.4 minutes.
When results differ
The overall index score does not predict every evaluation.
- FrontierCode 2.0 score: 49.1%
- Average task time: 17.9 minutes
- FrontierCode 2.0 score: 48.8%
- Average task time: 46.4 minutes
Use the estimates carefully
A real request’s bill depends on its input and output token use.
- Benchmark averages are not fixed per-task fees.
- Actual cost and quality can vary with the task and workload.
- Compare API use, completion time, and output quality on your own tasks.
The Cost of More Model Effort
The figures put a practical trade-off behind the model’s headline ranking: a business may pay substantially more and wait longer for a modest improvement on one benchmark. High effort scored 62 for $9.08, while Max scored 65 for $26.54. That makes High a potentially more economical choice for work where the benchmark is a useful guide, while Max may suit tasks where a small score improvement is worth the extra cost and time.
These are not fixed prices for every request. Actual API charges depend on a task’s input and output token use, and the index’s average may not predict a company’s workload. The results also show that more effort does not guarantee a higher result on every evaluation: High slightly outscored Max on FrontierCode 2.0. Teams should treat the published costs as a comparison point, not a quote for a particular project.
Popular right now
KitchenAid Classic Series 4.5-Quart Tilt-Head Stand Mixer, Onyx Black
Dyson V8 Cyclone Cordless Vacuum
Ninja CREAMi Scoop & Swirl 13-in-1 Ice Cream and Soft Serve Maker, NC701
As an affiliate, we earn on qualifying purchases.
Fable 5.2 and Its Effort Settings
Anthropic released Claude Fable 5.2 on November 9, alongside Claude Mythos 5.2. The source report describes Mythos as the same underlying model with fewer safety restrictions, offered only to organizations accepted into Anthropic’s Cyber or Life Sciences Verification Programs. Fable is available to users generally and has safeguards aimed at blocking high-risk cybersecurity work and most research-biology tasks. Anthropic says blocked queries are answered by an older model instead.
Artificial Analysis placed Fable 5.2 at the top of its Intelligence Index at Max effort, with the Xhigh and High entries close behind. Anthropic says the new models generate answers up to 30% faster than their 5.1 versions and lists the same standard token prices for each Fable effort level. An optional fast mode, according to the source material, costs $16 per million input tokens and $80 per million output tokens. Anthropic recommends starting with its Opus model for most coding and knowledge work and moving to Fable when a task calls for it.
In the same Artificial Analysis comparison, Opus 5.5 at Max scored 59 for $10.04 per task; Fable 5.2 at High scored 62 for $9.08. That comparison is specific to the index and its settings. It does not establish that Fable will be cheaper or perform better on every user’s workload.
What the Benchmark Cannot Predict
The published figures do not show what a specific customer will pay for a particular request. Task complexity, prompt length and generated output can affect token use, and the source material does not provide a distribution of costs across individual tasks or customer workloads. The benchmark’s average should not be read as a fixed per-task fee.
Scores are shown as rounded values, so identical displayed scores at High and Xhigh do not establish that the underlying results were exactly equal. The supplied material also does not explain the full uncertainty around the scores or how closely the index reflects each type of work. Anthropic’s reported demonstrations and claims about speed are company statements, not independent confirmation of typical performance. The cited Artificial Analysis results are dated November 10, 2026; later model or pricing changes are not covered.
Choosing a Setting for Workloads
Businesses can use the five published settings as a starting point, then compare API usage, completion time and output quality on their own tasks. Medium is the default; the benchmark suggests High may offer a different cost-to-score balance than Max, but local testing is needed before changing a production workflow. Teams should also check current Anthropic pricing and model guidance, since the figures here reflect the rates and benchmark results reported on November 10, 2026.
No further benchmark release or pricing change is specified in the source material. For now, the central decision remains workload-specific: whether the additional time and cost at higher effort levels produce a useful improvement for the task at hand.
Key Questions
How much does a Claude Fable 5.2 task cost?
Artificial Analysis estimates an average of $3.10 per task at Low, rising to $26.54 at Max. These are averages for tasks in its Intelligence Index, not fixed prices for every request.
Which effort level offers the best score for the money?
The benchmark does not establish a single best choice for every user. High scores 62 for $9.08 per task, while Max scores 65 for $26.54. Businesses should compare results on their own work before choosing a setting.
Why does the cost change if token prices stay the same?
Anthropic’s listed per-token rates are the same across Fable effort levels, but token use rises with effort in the reported test. Artificial Analysis recorded 34.9 million output tokens across the index at Low and 275 million at Max.
Is Max always more accurate than High?
No. Max received a higher score on the overall Intelligence Index, but on Artificial Analysis’s FrontierCode 2.0 test, High scored 49.1% and Max 48.8%. Results vary by evaluation.
Are these prices guaranteed for everyday use?
No. They are benchmark-based average API costs. A real request’s bill depends on its input and output token use, and may differ from the index estimate.
Source: ThorstenMeyerAI.com
