Claude Fable 5.2 Posts the Highest Score Ever on the Artificial Analysis Intelligence Index
Claude Fable 5.2 Scores 65 on Artificial Analysis, Highest So Far
Artificial Analysis benchmarked Claude Fable 5.2 (max) at 65 on its Intelligence Index v5.0.1, the top result recorded so far. The index aggregates nine evaluations, including AA-Briefcase, AutomationBench-AA, Terminal-Bench 4.0, FrontierCode 2.0, AA-Omniscience, GDP.pdf, HLE-Diamond, CritPt v2, and Terminal-Bench-Science. The analysis also breaks down cost per task, output tokens, pricing, and context window, though no public API is available for the model.
AA-Omniscience v1.1 Net Score measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer.
- rgbrenner
Forget about the scores... look at the Cost Per Task. Fable 5.1 was already the most expensive by a good margin, and now Fable 5.2 doubles it! This is with the price cut.
The output tokens per task just exploded. From 118k tokens (46k Answer/71k Reasoning) in 5.1 to 274k tokens (81k Answer/192k Reasoning) in 5.2.
- oliwary
Cool! It looks like Fable 5.2 (especially low and medium) also sit on the pareto frontier between time per task and intelligence: https://artificialanalysis.ai/models?speed=output-speed-over...
I am always interested in this comparison since it shows how quickly you can go back and forth with a model, capturing both tokens/task and tokens/second. The new models seem to be meaningfully better here.
- brachkow
It's crazy how expensive Anthropic's model is compared to pretty capable DeepSeek-Flash
While they made, hands down, the best models, their 100x price for few percent better is completely made up, and I think it will end up bad for them
- djacksondev
To me Fable's role was always as a planner or plan reviewer, rather than something individuals or organizations would use as the primary driver. In that case, I think these improvements seem worthwhile in exchange for the cost? Ideally in that role, you want the best performing model.
Not sure if the current rise means it's cost prohibitive to use in that form too.
- schnitzelstoat
That seems like an insane increase in cost for a modest intelligence increase.
But perhaps the performance on real-world tasks will be more noticeable than on benchmarks.
- t0md4n
8 points higher than Astra for 4.2x the cost and 5x the output tokens. Is it worth it?
Edit: high seems better than max
- gizmodo59
https://artificialanalysis.ai/#intelligence-efficiency-tabs this is the chart that matters for most of poor developers and companies. Pareto for a given intelligence index and looks like only openai is across the quadrant with muse at the bottom. They have really cooked efficient models!
- GMoromisato
Anecdotal, but I have the model write a podcast script of two reviewers talking about a book, one chapter at a time.
Fable (5.1 and 5.2) consistently write a much shorter script (2.5 pages) compared to Opus 5.5 (4.5 pages).
Fable 5.2: https://claude.ai/share/274a96a8-c8a5-4564-9ca1-e914edffc722
Opus 5.5: https://claude.ai/share/da8ab6bd-0987-47dc-9c26-68ebbfb25781