🔍 Read the full analysis: A Practical Look At Claude Fable 5.2 Vs. Gemini 4 For Work on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Artificial Analysis scores Claude Fable 5.2 highest on its Intelligence Index, but Gemini 4 Carbon leads several individual work tests and costs less per task. Fable scores better on some coding and long-document evaluations, while Carbon has a much lower reported hallucination rate when it lacks an answer. The results do not establish one best model for every workplace task.
A comparison published by ThorstenMeyerAI.com on November 17 finds that Anthropic’s Claude Fable 5.2 leads Google’s Gemini 4 models on Artificial Analysis’s overall Intelligence Index, while Gemini 4 Carbon leads several individual work tests and costs less per task. The results offer businesses no single winner: performance varies by assignment, and reported prices and reliability measures differ substantially.
On the Artificial Analysis Intelligence Index, Fable 5.2 scored 65 at Max effort and 62 at High. Gemini 4 Carbon scored 60 at High, while Gemini 4 Hydrogen scored 54 at its highest listed setting and Gemini 4 Argon scored 50 at High. These are results from one benchmark index, not a measure of performance in every company’s workflows.
Individual evaluations split the comparison. Fable 5.2 (High) scored 49.1% on FrontierCode-AA 2.0, which tests agentic coding in open-source repositories, compared with Carbon’s 40.6%. Fable also led on the AA-LCR long-context reasoning test, 83.3% to 79.3%. Carbon scored higher on Terminal-Bench 4.0, professional-document test GDP.pdf and business-workflow test AutomationBench-AA: respectively 65.2% versus 59.6%, 28.4% versus 21.8%, and 77.8% versus 71.4%.
The article’s Artificial Analysis figures put the cost of one Intelligence Index task at $26.54 for Fable at Max and $9.08 at High, compared with $5.48 for Carbon, $4.97 for Argon and $3.16 for Hydrogen. Carbon and Argon costs use Google’s introductory pricing through November 30; the listed prices double on December 1. The source cautions that token rates alone do not capture how many tokens a model uses while reasoning.
Choosing Models by Workload and Cost
For organizations comparing AI tools, the results point to a task-by-task procurement decision, not a straightforward ranking. A team focused on coding in repositories or long-document reasoning may value Fable’s stronger results on those tests. Teams doing terminal work, professional-document tasks or business-software automation may find Carbon’s benchmark leads more relevant.
Price also changes the trade-off. The comparison reports lower per-task costs for Gemini 4, although the benchmark’s task-cost figure is not a universal estimate of a company’s bill. Actual spending can depend on the prompt, amount of reasoning, model settings and workload. Organizations should test representative tasks and review error rates alongside cost before choosing a default model.
Popular right now

De'Longhi Classic Signature Espresso Machine with Adjustable Milk Frother

Dyson Airwrap Co-anda2x Multi-Styler and Dryer (2026 Model)

Hisense E7 Cinema Series 55-Inch Hi-QLED Mini-LED 4K UHD Smart Fire TV
As an affiliate, we earn on qualifying purchases.
November Releases and Model Tiers
Anthropic released Claude Fable 5.2 on November 9. Google opened Gemini 4 Carbon to paying users on November 10 and made the faster Gemini 4 Hydrogen available to everyone, including through Google Search, on November 12. Gemini 4 is a family: Google describes Carbon as its most capable reasoning model and Argon as an efficient option for most work. Hydrogen is positioned for rapid execution across agentic workflows.
The source says Fable 5.2 and all three Gemini 4 models have one-million-token context windows, though their supported inputs vary. Fable reads text and images; Carbon and Hydrogen also accept speech and video, according to Artificial Analysis. Anthropic’s listed API rates for Fable are $8 per million input tokens and $40 per million output tokens. Google listed Carbon and Argon at $2 and $10 through November 30, then $4 and $20; Hydrogen is listed at $1.50 and $7.50.
The comparison was researched and drafted with AI assistance, including an Anthropic model. Its author disclosed that conflict of interest and said neither Anthropic, Google nor Artificial Analysis reviewed or sponsored the article. The benchmark results and prices are attributed to the named sources and were accessed on November 17.
““the most ambitious, open-ended technical work””
— Anthropic, describing Claude Fable 5.2
Benchmark Limits and Error Rates
The reported scores do not establish how either model will perform on a particular employer’s data, software or instructions. The source gives selected benchmark results, and some tests show low absolute scores: on GDP.pdf, for example, both models scored below 30%. Benchmark leaders may not translate directly into faster, more accurate or cheaper work in everyday use.
The article also reports a difference in uncertainty handling: Fable answered accurately on 77.2% of questions, compared with Carbon’s 58.3%, but its hallucination rate when it did not know an answer was 58.0%, versus 11.4% for Carbon. The supplied material does not detail the test protocol or define every denominator behind these figures, so they should be read as results from the cited evaluation rather than universal rates of workplace errors.
Other comparisons carry limits. The source says some vendor-reported evaluations cannot be independently repeated, and the Epoch AI Capabilities Index estimates for Fable and Carbon have overlapping uncertainty ranges. Google’s introductory Carbon and Argon prices also change on December 1. The comparison does not provide an independent cost audit across real customer workloads.
Recheck Prices After November
The clearest near-term change is Google’s scheduled price adjustment: Carbon and Argon API rates in the comparison are set to double on December 1 from their introductory levels. Businesses using those models will need to recalculate costs against their actual token use and task volume after the change.
For teams deciding now, the next practical step is to run both models on the same representative tasks, with the intended settings and review criteria. That can show whether benchmark differences matter to the organization’s work and how often outputs need correction. The cited comparison provides no announced future benchmark milestone or independent head-to-head workplace trial.
Key Questions
Which model scored highest on the overall index?
Claude Fable 5.2 at Max effort scored 65 on the Artificial Analysis Intelligence Index accessed November 17, 2026. Fable at High scored 62, and Gemini 4 Carbon at High scored 60.
Is Gemini 4 Carbon better for every work task?
No. Carbon led the cited terminal, professional-document and business-workflow tests, while Fable led the open-source repository coding and long-context reasoning tests. The results vary by benchmark.
Which model costs less per benchmark task?
The reported Artificial Analysis task costs were lower for the Gemini 4 models: $5.48 for Carbon, $4.97 for Argon and $3.16 for Hydrogen, compared with $9.08 for Fable at High and $26.54 at Max. Carbon and Argon figures use introductory prices that the source says double on December 1.
What do the reported reliability figures say?
In the cited evaluation, Fable had higher answer accuracy, at 77.2% versus Carbon’s 58.3%. When the models did not know an answer, the reported hallucination rates were 58.0% for Fable and 11.4% for Carbon. These are evaluation-specific figures, not guaranteed workplace error rates.
Does the comparison prove which model a company should use?
No. It reports benchmark scores and prices, but does not test every company’s tasks or provide a universal cost estimate. Organizations would need to compare the models on representative work and account for their own usage and review needs.
Source: ThorstenMeyerAI.com
