🔍 Read the full analysis: Which Claude Fable 5.2 Effort Level Fits Your Budget Per Task? on ThorstenMeyerAI.com
Get the little things that make your day 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
Artificial Analysis’ November 10 comparison puts Claude Fable 5.2’s benchmark task costs at $3.10 for Low effort and $26.54 for Max, with scores of 54 and 65, respectively. High scores 62 for $9.08 per task, while Xhigh costs $14.83 for the same rounded score; the figures come from a defined benchmark and are not a forecast of every company’s bill.
Claude Fable 5.2 is available with five effort settings, and a November 10 comparison by Artificial Analysis shows benchmark task costs ranging from $3.10 at Low to $26.54 at Max. Scores on the organization’s Intelligence Index rise from 54 to 65 across those settings, giving businesses a concrete cost-and-performance comparison—but not a guarantee of what a typical task will cost or score.
Artificial Analysis tested Fable 5.2 at Low, Medium, High, Xhigh and Max on its Intelligence Index, which covers nine evaluations in areas including agentic coding, business documents, science and knowledge. The reported average API cost per task was $3.10 at Low, $3.80 at Medium, $9.08 at High, $14.83 at Xhigh and $26.54 at Max. The corresponding scores were 54, 56, 62, 62 and 65.
The cost figures are specific to that test. Anthropic’s stated list prices are $8 per million input tokens and $40 per million output tokens at every effort setting. What varies is how many tokens the model uses while working. Artificial Analysis reports total output-token use across the index rising from 34.9 million at Low to 275 million at Max. Average time per task also grows, from 6.5 minutes at Low to 46.4 minutes at Max.
The comparison shows the largest score increase between Medium and High: six points for $5.28 more per benchmark task. Xhigh costs more than High but has the same rounded score of 62 in the published table. Max adds three points over High while costing nearly three times as much. On one separate agentic coding evaluation, Artificial Analysis reported High at 49.1% and Max at 48.8%, a result that cautions against assuming more effort will improve every task.
Which Claude Fable 5.2 Effort Level Fits Your Budget Per Task?
Artificial Analysis’ November 10 benchmark shows how task cost, score and time change across five effort settings. Treat its average costs as comparison points, not quotes for your own workload.
Cost rises with effort
Benchmark averages show a steep increase in cost at the higher settings.
Compare the trade-offs
Scores are from an index of nine evaluations; they do not predict every task outcome.
| Effort | Index score | Average time per task | What stands out |
|---|---|---|---|
| Low | 54 | 6.5 minutes | Lowest cost |
| Medium | 56 | Not provided | Low benchmark spend |
| High | 62 | Not provided | Same rounded score as Xhigh |
| Xhigh | 62 | Not provided | Higher cost than High |
| Max | 65 | 46.4 minutes | Highest index score and cost |
Choose by workload
The right setting depends on the value of extra performance, the time available and your own task mix.
- Consider Low or Medium for routine work and tighter task budgets.
- Check whether the score and completion quality meet your needs.
- Consider High or Max when added performance could justify more spend and waiting.
- Test Xhigh against High: both scored 62, but Xhigh cost more.
What the figures mean
Anthropic’s stated token prices are unchanged across effort settings; usage varies with how many tokens the model uses.
Use a representative mix from your own workflows.
Compare cost, completion quality and elapsed time by setting.
Reserve higher effort for work where it improves results enough to justify the trade-off.
Budgeting for More Model Effort
The figures give organizations a way to think about effort as a task-level spending choice, rather than treating the model’s headline ranking as the only measure that matters. For repeated work, the difference between a benchmark average of $3.80 at Medium and $26.54 at Max could affect operating budgets, response times and how teams decide which requests deserve more computation.
They also show why the highest score is not automatically the best business setting. High recorded 62 for $9.08 per task, while Max recorded 65 for $26.54 and took much longer on average. A team doing work where a small score difference has real value may prefer Max; for routine requests or workflows sensitive to waiting time, a lower level could be a better fit. That is an interpretation of the benchmark, not a universal recommendation.
Real spending will depend on a company’s prompts, task mix, token use and API settings. The test’s average cost is useful for comparing settings under one evaluation, but it should not be read as a fixed quote for every customer request.
Products Worth Considering
LEVOIT Vital 200S-P Smart Air Purifier for Large Rooms up to 1875 ft², Quiet 27 dB Operation, AHAM VERIFIDE Certified, Air Quality Monitor,
Samsung 43-Inch M70H Series 4K Mini LED Smart TV (2026)
Dyson Airwrap i.d. Multi-Styler for Straight and Wavy Hair
As an affiliate, we earn on qualifying purchases.
What the Benchmark Measures
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, replacing their respective 5.1 versions, according to the source material. Fable is the model offered to users generally, with safeguards for high-risk cybersecurity and much research biology work. Mythos is described as the same underlying model without those safeguards and is restricted to organizations accepted into Anthropic’s Cyber or Life Sciences Verification Programs.
Artificial Analysis ranked Fable 5.2 at Max first among 74 models on its Intelligence Index in the comparison cited here. Its next listed Fable settings were also near the top: Xhigh and High each scored 62, with published ranks of second and third. Rankings and scores describe performance on that index, not every type of work a model may be asked to do.
Anthropic recommends starting with Claude Opus 5.5 for most coding and knowledge-work tasks, according to the supplied material, and moving to Fable when a task calls for it. In Artificial Analysis’ comparison, Opus 5.5 at Max scored 59 at $10.04 per index task, while Fable 5.2 at High scored 62 at $9.08. Those results suggest a potentially useful benchmark comparison, but they do not establish that Fable will be cheaper or more effective for every workload.
Limits of the Cost Comparison
The supplied figures do not establish how often businesses will see the benchmark’s average cost per task in their own use. Actual bills depend on prompts, input and output token counts, task design and the effort setting selected. Artificial Analysis’ index combines nine evaluations, so its average does not describe any one customer’s workload.
The source material also does not provide enough detail to independently assess every benchmark methodology choice or reproduce the results. Anthropic’s claims about faster answers, safeguards and achievements from prerelease testing are company statements; the described benchmark scores do not independently verify those claims. The reported differences should be treated as a snapshot of the cited comparison, not proof that one setting will outperform another in all circumstances.
Test Settings on Real Work
Businesses considering Fable 5.2 can compare its settings on a representative sample of their own tasks, tracking cost, completion quality and elapsed time rather than relying only on an overall index score. A useful evaluation would separate routine requests from tasks where additional reasoning could justify higher cost, and check whether High or Max changes the outcome enough to warrant the added time and spend.
The available report is a November 10 benchmark snapshot. Further comparisons from Artificial Analysis, additional real-world usage data and clearer workload-specific cost estimates could help show how closely these averages match customer experience. Until then, the practical choice depends on each organization’s own tasks and budget.
Key Questions
How much does Claude Fable 5.2 cost per task?
In Artificial Analysis’ Intelligence Index comparison, average task costs ranged from $3.10 at Low to $26.54 at Max. These are benchmark averages, not a fixed price for every user request.
Which effort level gives the best score for its cost?
The benchmark does not declare one setting best for every use. High scored 62 for $9.08 per task, while Xhigh also scored 62, rounded, for $14.83. Whether the extra cost is justified depends on the task and the value of additional performance.
Does Max effort always perform better than High?
No. Max scored higher on the Intelligence Index, at 65 compared with High’s 62. But Artificial Analysis reported a slightly higher result for High than Max on its FrontierCode 2.0 coding test: 49.1% versus 48.8%.
Are the benchmark task costs what a company will pay?
Not necessarily. They are average API costs within the cited benchmark. A company’s bill will vary with its workload and token use. Anthropic’s stated list prices are $8 per million input tokens and $40 per million output tokens across the five effort settings.
Source: ThorstenMeyerAI.com
