GPT-6.2 Astra and Sol score lower on transparency than Claude
Original title: GPT-6.2 Astra and Sol get to the point faster but score lower on transparency than Claude
Astra's average reply shrank from 352 to 210 words, and its admits-uncertainty score fell from 96 to 24.
Artificial Analysis compared OpenAI's GPT-6.2 Astra and Sol with Anthropic's Claude Fable 5.2 and Opus 5.5 on AA-PersonalityBench. Both OpenAI models give shorter, less structured answers and rank lower on admitting uncertainty and transparency. Claude leads those indexes, while Opus 5.5 leads informativeness. The firm says the scores describe communication style, not answer accuracy.
Unconfirmed: Personality scores are relative ranks from Artificial Analysis's AI judges, not a measure of factual accuracy.
- Astra headings per answer fell from 1.99 to 0.11; Sol from 1.85 to 0.25.
- Organization Index fell from 70 to 11 for Astra and from 86 to 27 for Sol.
- Transparency scores were 81 for Fable 5.2 and 90 for Opus 5.5, versus 41 and 57.
- The bench uses 900 English questions judged by a panel of three AI models.
- Scores from 1 to 100 are relative ranks among tested models, not answer percentages.
- Organizations
- Artificial Analysis
From the original article
3 min read at the-decoder.comOpenAI's GPT-6. 2 Astra and Sol give much shorter, more direct answers than their predecessors, but they are also less willing to admit uncertainty. Anthropic's Claude models score well ahead on transparency.
By Tomislav Bezmalinović · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.