🔍 Read the full analysis: Claude Fable 5.2 Costs Per Task: What To Expect At Each Effort Level on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Artificial Analysis reports that Claude Fable 5.2’s average cost per task on its Intelligence Index ranges from $3.10 at Low effort to $26.54 at Max. The highest score is 65 at Max, but High scores 62 for $9.08; the figures reflect one benchmark’s API costs, not a guaranteed price for ordinary work.
Artificial Analysis estimates that Claude Fable 5.2 costs an average of $3.10 to $26.54 per task on its Intelligence Index, depending on the selected effort level. The results, published after Anthropic announced the model on November 9, show a score of 54 at Low effort and 65 at Max, giving businesses a benchmark-based view of how additional model reasoning may affect cost and speed.
Artificial Analysis tested five effort settings—Low, Medium, High, Xhigh and Max—on its Intelligence Index, which covers nine evaluations in areas including agentic coding, business documents, science and general knowledge. Its reported “cost per task” is the average API cost for one task in that test. The figures are not a fixed charge for every user request or a forecast of a company’s total bill.
The benchmark lists Medium as the default in Claude apps and Claude Code. It scored 56 at an average $3.80 per task, compared with 54 and $3.10 at Low. High scored 62 at $9.08, Xhigh scored 62 at $14.83, and Max scored 65 at $26.54. Average task times ranged from 6.5 minutes at Low to 46.4 minutes at Max.
Anthropic says the standard API price is $8 per million input tokens and $40 per million output tokens at every effort level. The cost difference in the benchmark comes from how many tokens the model uses while working: Artificial Analysis reports 34.9 million output tokens across the index at Low and 275 million at Max. Anthropic also offers an optional fast mode at higher token rates, according to the source material.
Claude Fable 5.2 costs by effort level
Artificial Analysis reports higher benchmark costs as effort increases. These averages reflect one test, not a guaranteed price for ordinary work.
Cost rises with effort
Scores, costs and times are averages from Artificial Analysis’s Intelligence Index.
| Effort | Score | Cost per task |
|---|---|---|
| Low | 54 | $3.10 |
| Medium | 56 | $3.80 |
| High | 62 | $9.08 |
| Xhigh | 62 | $14.83 |
| Max | 65 | $26.54 |
What the extra spend buys
- Score: 62
- Average cost: $9.08 per task
- FrontierCode 2.0 score: 49.1%
- Score: 65
- Average cost: $26.54 per task
- Average task time: 46.4 minutes
Why benchmark costs differ
Anthropic’s standard API rates are reported as the same across effort levels; token use drives the benchmark’s cost gap.
$8 per million input tokens and $40 per million output tokens.
Artificial Analysis reports 34.9 million output tokens at Low and 275 million at Max across the index.
More tokens used in the test corresponded to higher average API cost per task.
Choose effort for the workflow
The benchmark does not show that more effort always improves results: High scored slightly above Max on the reported coding test.
- Try representative tasks at Medium, High and Max.
- Track output quality, response time and token use.
- Check current prices, access and safety behavior with Anthropic.
The Price of More Reasoning
The results give teams a reason to choose effort settings by task rather than defaulting to the most expensive option. In this test, moving from Medium to High added six score points for $5.28 more per task. Moving from High to Xhigh cost another $5.75, while both settings had the same rounded score of 62. Max cost nearly three times as much as High for three additional points.
Time is part of the trade-off. A task averaging 46.4 minutes at Max may be a poor fit for workflows where users need a quick response, even if the higher score is useful for difficult work. The figures are benchmark averages, though, and do not establish how much a particular business task will cost or how much time it will take.
Artificial Analysis also reports that on its FrontierCode 2.0 coding test, Fable 5.2 at High scored 49.1%, slightly above Max at 48.8%. That result cautions against assuming that more effort always improves performance. It is one test result, not evidence that High is better for every coding task.
Products Worth Considering
LG UltraGear 25G590B-B 25-Inch Full HD 1000Hz Gaming Monitor
Logitech MX Ergo S Advanced Wireless Trackball Mouse USB-C – Graphite
Apple AirPods Pro (2nd Generation) Wireless Earbuds with USB‑C Charging, Up to 2X More Active Noise Cancelling, Transparency Mode, Adaptive
As an affiliate, we earn on qualifying purchases.
How the Effort Test Works
Anthropic announced Claude Fable 5.2 and Claude Mythos 5.2 on November 9, replacing the previous 5.1 versions. The source material describes Fable as available to users with safeguards for some high-risk cybersecurity and biology requests. Mythos is described as the same underlying model without some of those safety restrictions, limited to organizations accepted into Anthropic’s verification programs for cyber or life sciences work. These access and safety details are Anthropic’s statements.
Anthropic says the new models produce answers up to 30% faster than the 5.1 versions. That is a company claim, not a timing result from the Artificial Analysis effort-level table. Anthropic recommends that most coding and knowledge-work users start with Claude Opus 5.5 and move to Fable only when a task calls for it, according to the source material.
On the same Artificial Analysis index, the supplied figures put Opus 5.5 at Max at 59 points and $10.04 per task, while Fable 5.2 at High scored 62 at $9.08. The comparison is useful as a reported benchmark snapshot, but does not settle which model is better for a particular workload. Rankings and costs depend on the test setup and can change as models or evaluations are updated.
“Start with Claude Opus 5.5 for most coding and knowledge work, and move up to Fable when the task needs it.”
— Anthropic
Limits of the Cost Estimates
The reported costs come from Artificial Analysis’s Intelligence Index, not a representative bill for every real-world use. The supplied material does not establish how the average changes across different prompt lengths, workloads, account arrangements or production usage. Individual costs will depend on token use and the task being run.
The source gives rounded scores and average test times but does not provide enough detail here to independently reproduce the full evaluation. It also does not show whether the reported scores or prices have since changed. Anthropic’s speed, safety and capability statements remain company claims unless independently verified; the benchmark figures apply only to the stated evaluation.
What Buyers Should Check
Teams considering Fable 5.2 can use the published results as a starting point, then test their own tasks at Medium, High and Max before setting a default. Tracking output quality, response time and token use on representative work will show whether a higher effort level justifies its additional cost in that workflow.
Anthropic’s recommendation is to begin with Opus 5.5 for most coding and knowledge work and use Fable when greater capability is needed. Buyers should also confirm current API prices, model access and safety behavior directly with Anthropic, since the figures in this report are dated to November 10, 2026 and the benchmark does not predict a company’s actual bill.
Key Questions
How much does Claude Fable 5.2 cost per task?
Artificial Analysis reports average costs on its Intelligence Index ranging from $3.10 at Low effort to $26.54 at Max. These are test averages, not fixed prices for all requests.
Which effort level offers the highest reported score?
Max scored 65 on the reported index, compared with 62 at High and Xhigh. Artificial Analysis lists Max at an average $26.54 per task.
Is High effort a cheaper alternative to Max?
In this test, High scored 62 for $9.08 per task, while Max scored 65 for $26.54. The benchmark does not show that High will suit every task, but it reports a substantially lower average cost.
Why do costs differ if token prices are the same?
Anthropic’s standard rates are reported as $8 per million input tokens and $40 per million output tokens at each effort level. Artificial Analysis attributes the different task costs to variation in token use, including tokens spent on reasoning.
Are these figures a guarantee of what my API bill will be?
No. They are average API costs for tasks in one benchmark. Actual charges depend on usage and token consumption, and the source material does not provide a forecast for any particular business workflow.
Source: ThorstenMeyerAI.com
