TL;DR
Artificial Analysis data available on November 17 puts Claude Fable 5.2 slightly ahead of Gemini 4 Carbon on its overall Intelligence Index, but results vary by task. Carbon leads several workplace tests and costs less per index task; Fable scores higher on repository coding, long-context reasoning and answer accuracy, while Carbon has a lower measured hallucination rate.
Claude Fable 5.2 and Google’s Gemini 4 models reached customers within days of one another in November, and independent benchmark results show no single model leading every work task. Artificial Analysis’s measurements put Fable 5.2 ahead of Gemini 4 Carbon on its overall index, while Carbon leads several workplace tests and costs less per measured task.
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
Anthropic released Claude Fable 5.2 on November 9. Google opened Gemini 4 Carbon to paying users the following day, then made the faster Gemini 4 Hydrogen available to everyone on November 12, including through Google Search. The source article dates its comparison to November 17 and cites model pages from Artificial Analysis.
At the High effort setting, Fable 5.2 scored 62 on the Intelligence Index, compared with 60 for Carbon. Fable led the reported tests for coding in open-source repositories, at 49.1% versus Carbon’s 40.6%, and long-context work, at 83.3% versus 79.3%. Carbon scored higher on terminal tasks (65.2% to 59.6%), documents (28.4% to 21.8%) and business software automation (77.8% to 71.4%).
Costs also differed in the reported comparison: an index task was priced at $9.08 for Fable 5.2 at High and $5.48 for Carbon. At Max effort, Fable scored 65 and cost $26.54 per task. The source says Carbon and Argon’s introductory prices double on December 1; Hydrogen’s reported task cost was $3.16. These task-cost estimates account for more than published token rates, since models can spend differing amounts of tokens while working.

Choosing a Model by Workload
The results matter to organizations deciding which model to use for coding, document work or automated business processes. A top score on an aggregate index does not predict the best result for every task: Fable led on repository coding and long context, while Carbon led on three reported workplace categories. Teams may get a more useful comparison by measuring their own workloads and tracking both output quality and cost.
The reported reliability measures add another consideration. Fable answered more questions correctly in the cited test, with 77.2% accuracy against Carbon’s 58.3%. But its reported hallucination rate was also higher: 58.0% for Fable versus 11.4% for Carbon. Those figures describe the benchmark cited by the article; they do not by themselves establish how either model will perform in a particular company’s workflow.
Price affects how often a model can be used and which jobs are economical to assign to it. Carbon’s and Argon’s listed introductory rates change on December 1, according to the source. Buyers comparing options should account for that scheduled change and the cost of a completed task, as well as the per-token price.
Products Worth Considering

Apple MacBook Air 15-inch (2026) with M5 chip, 16GB memory, 512GB SSD — Midnight

Logitech MX Creative Console with 9 Customizable Keys, Graphite

Samsung Galaxy Tab A11+ Kids Edition, 128GB, Royal Blue
As an affiliate, we earn on qualifying purchases.
A Three-Model Gemini Rollout
Google’s Gemini 4 is a family rather than a single model. The source describes Carbon as Google’s most capable reasoning model, Argon as its efficient frontier model, and Hydrogen as the faster option for agentic workflows. Google recommends Argon for most software engineering and knowledge work, according to the source article. All three are listed with a one-million-token context window.
Anthropic describes Fable 5.2 as a model for ambitious, open-ended technical work. It offers five effort settings, a one-million-token context window and up to 128,000 output tokens. The source reports that Fable accepts text and images; Carbon and Hydrogen also accept speech and video, while Argon accepts text and images.
Pricing depends on the model and date. The source lists Fable at $8 per million input tokens and $40 per million output tokens, with fast mode at twice those rates. Carbon and Argon are listed at $2 and $10 through November 30, then $4 and $20 from December 1. Hydrogen is listed at $1.50 and $7.50. Artificial Analysis’s task-cost figures offer a separate comparison, since token prices alone do not show how many tokens a model uses to complete a task.
Limits of the Benchmark Results
The figures are benchmark results accessed November 17, not guarantees of performance on a reader’s own data or software. The supplied source names the tests and reports scores, but its available text does not include the complete explanation of test methods or results for every model and category. It also does not establish how the reported hallucination rates translate into errors in any specific workplace setting.
The comparison’s prices are date-sensitive: the source says Carbon and Argon introductory rates rise on December 1. Actual spending will depend on usage, model settings and tokens consumed. The material also does not give a direct, complete comparison of every Gemini 4 model against Fable 5.2 across every listed benchmark.
Recheck Prices and Test Tasks
Organizations weighing the models can compare them on representative tasks, then review accuracy, error rates, completion time and cost for each workflow. Buyers using Carbon or Argon should account for the scheduled December 1 price change identified in the source. Further benchmark results or updated pricing could change the comparison, but the supplied material does not identify a forthcoming test release or other next milestone.
Key Questions
Which model scored highest on the overall index?
Claude Fable 5.2 scored 62 at High effort, compared with 60 for Gemini 4 Carbon, in the Artificial Analysis measurements cited by the source and accessed November 17, 2026.
Which model is better for workplace tasks?
It depends on the task. Fable led the reported repository-coding and long-context tests. Carbon led terminal work, documents and business software automation.
Which model costs less per index task?
The reported estimate was $5.48 for Gemini 4 Carbon and $9.08 for Fable 5.2 at High effort. The source says Carbon’s introductory price doubles on December 1, so buyers should check the applicable rate.
What do the reported reliability figures show?
Fable had higher answer accuracy in the cited test, 77.2% versus Carbon’s 58.3%. Carbon had the lower reported hallucination rate, 11.4% versus 58.0%. These are benchmark measurements, not a prediction for every use case.
Source: Thorsten Meyer AI
