TL;DR
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
Artificial Analysis reports that Claude Fable 5.2 costs an average of $3.10 per task at Low effort and $26.54 at Max on its Intelligence Index. The Max setting scores 65, compared with 62 at High for $9.08, while the underlying API token rates are the same at every effort level.
Artificial Analysis estimates that Claude Fable 5.2 costs $3.10 to $26.54 per task across five effort settings on its Intelligence Index, with scores rising from 54 at Low to 65 at Max. The figures give businesses a way to compare the additional benchmark performance against higher average API bills and longer task times.
On the index, Low effort averages $3.10 per task and takes 6.5 minutes, while Medium, the default setting in Claude apps and Claude Code, averages $3.80 and 8.3 minutes. High averages $9.08 per task and 17.9 minutes; Xhigh costs $14.83 and takes 26.4 minutes; Max costs $26.54 and takes 46.4 minutes. Artificial Analysis describes these as average API costs for a task in its test suite, not fixed prices for every real-world request.
The corresponding scores are 54 at Low, 56 at Medium, 62 at High, 62 at Xhigh and 65 at Max. The largest score increase in the reported figures comes between Medium and High: six points for an additional $5.28 per task. Xhigh costs about 63% more than High but receives the same rounded score. Max adds three points over High while costing nearly three times as much.
Anthropic lists the same token rates for all effort settings: $8 per million input tokens and $40 per million output tokens. The difference in tested bills reflects token use: 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 $16 per million input tokens and $80 per million output tokens, according to the source material.
Claude Fable 5.2 cost per task at each effort level
Higher effort brings higher benchmark scores, bills and task times. These are Artificial Analysis averages for its Intelligence Index, not fixed prices for business requests.
Cost, score and time rise with effort
The figures compare five benchmark settings; higher scores do not guarantee better results for every job.
| Effort | Cost per task | Score | Task time |
|---|---|---|---|
| Low | $3.10 | 54 | 6.5 min |
| Medium | $3.80 | 56 | 8.3 min |
| High | $9.08 | 62 | 17.9 min |
| Xhigh | $14.83 | 62 | 26.4 min |
| Max | $26.54 | 65 | 46.4 min |
Where the extra spend goes
Anthropic lists the same token rates across effort levels; the benchmark bills differ with token use.
- $8 per million input tokens
- $40 per million output tokens
- Optional fast mode: $16 input and $80 output per million tokens
- Low: 34.9 million output tokens across the index
- Max: 275 million output tokens across the index
The biggest score gain is at High
Moving from Medium to High adds six score points for $5.28 more per task. Xhigh costs about 63% more than High but has the same rounded score; Max adds three points over High at nearly three times the cost.
Medium is the default in Claude apps and Claude Code. Anthropic recommends Opus 5.5 for most coding and knowledge work.
Check whether the benchmark score gain at higher effort is worth the added cost and wait time for your tasks.
Actual bills vary with prompts, outputs, task type and reasoning needs; benchmark averages are not guaranteed prices.
The Cost of Higher Effort
The estimates show that selecting an effort level can change the expected bill and wait time substantially, even when the listed per-token rates stay constant. For teams running many tasks, moving from Medium to High raises the reported average from $3.80 to $9.08 per task; using Max raises it to $26.54 and stretches the average task time to 46.4 minutes.
Those costs matter because a higher benchmark score does not guarantee a better result for every job. Artificial Analysis reports that Fable 5.2 at High scored 49.1% on its FrontierCode 2.0 agentic coding test, slightly above the 48.8% reported at Max. The index and coding figures are test results, not predictions of performance on a particular company’s workload.
The comparison in the source material also puts Fable 5.2 High at 62 points and $9.08 per task against Claude Opus 5.5 at Max at 59 points and $10.04. Anthropic separately said Fable 5.2 at High scored 45.3% on FrontierCode 2.0, near Opus 5.5’s reported 45.5%, at roughly half the cost per task. These comparisons are attributed claims and should be read within the tests and conditions used.
Products Worth Considering
LG UltraGear 32GX850A-B 32-inch 4K UHD OLED Gaming Monitor
LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz
LG UltraGear 25G590B-B 25-Inch Full HD 1000Hz Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Fable 5.2’s Effort Settings
Anthropic announced Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026, according to the supplied material. It describes Mythos as the same underlying model with fewer safety restrictions, available only to organizations accepted into Anthropic’s Cyber or Life Sciences Verification Programs. Fable is presented as the version available to general users, with safeguards for high-risk cybersecurity and many research biology requests.
Artificial Analysis evaluated five Fable effort levels on its Intelligence Index, which the source describes as nine evaluations spanning agentic coding, business documents, science and knowledge. The ranking and task costs cited here were accessed November 10. The source says Fable 5.2 ranked first among 74 models at Max, with High and Xhigh also near the top; rankings and scores can change as tests and model entries are updated.
Anthropic says Fable 5.2 produces answers up to 30% faster than its 5.1 counterpart. The source also reports Anthropic’s recommendation to start with Claude Opus 5.5 for most coding and knowledge work, moving to Fable when a task calls for it. That guidance is the company’s position, not an independent finding.
“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 listed amounts are averages from Artificial Analysis’s benchmark tasks. Actual costs will vary with prompt length, output length, task type and how much reasoning a request uses; the figures do not establish what a particular business will pay for its own work.
The supplied source does not provide the full test methodology, cost calculation details or uncertainty ranges for each estimate. Scores are rounded as shown by Artificial Analysis, so equal displayed scores at High and Xhigh do not establish identical underlying performance. Anthropic’s claims about speed, safeguards and coding results have not been independently verified in the supplied material.
The source also characterizes Fable and Mythos as the same underlying model with different safety controls, but does not provide technical details that would independently establish how the two deployments differ. Availability, access terms and the practical effect of fallback responses to blocked requests may depend on Anthropic’s implementation.
Choosing a Setting for Workloads
Businesses considering Fable 5.2 can compare its settings against their own tasks, checking both output quality and cost rather than relying on the index alone. The supplied material identifies Medium as the default in Claude apps and Claude Code, while Anthropic recommends Opus 5.5 as a starting point for most coding and knowledge work.
Further independent evaluations, fuller methodology and real-world usage data would help clarify whether the score gains at higher effort levels justify their added cost and delay for specific workloads. Artificial Analysis’s published figures may also change as it updates its model pages and tests.
Key Questions
How much does Claude Fable 5.2 cost per task?
Artificial Analysis reports average costs from $3.10 at Low to $26.54 at Max on its Intelligence Index. These are benchmark task averages, not a guaranteed price for every request.
What is the default effort level?
The supplied source says Medium is the default in Claude apps and Claude Code. Artificial Analysis estimates its index tasks at $3.80 each on average at that setting.
Does Max effort provide the best score?
Among the five settings in the reported Intelligence Index results, Max scores highest at 65. It costs $26.54 per task on average, compared with $9.08 at High, which scores 62.
Why do costs differ if token rates are the same?
Anthropic lists the same per-token rates for every effort level. The benchmark costs differ because the settings use different amounts of tokens; Artificial Analysis reports 34.9 million output tokens across the index at Low and 275 million at Max.
Are the published task costs guaranteed for business use?
No. They are Artificial Analysis averages for its test suite. A company’s actual bill depends on its prompts, outputs and workload.
Source: Thorsten Meyer AI
