{"slug": "new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception", "title": "New benchmark confirms AI models still perform poorly at visual perception", "summary": "Moonshot AI's PerceptionBench benchmark shows that no frontier multimodal AI model reaches 60 percent accuracy in visual perception tasks, with GPT-5.6 Sol leading by a narrow margin. The benchmark separates visual perception from logical reasoning and reveals that many supposed reasoning errors actually occur during the image-reading stage.", "body_md": "Moonshot AI's PerceptionBench tests how well multimodal AI models can actually \"see,\" separate from logical reasoning. No frontier model reaches 60 percent accuracy, and GPT-5.6 Sol leads by a narrow margin. Many supposed reasoning errors actually happen as early as the image-reading stage.\n\nThe article [New benchmark confirms AI models still perform poorly at visual perception](https://the-decoder.com/new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception/) appeared first on [The Decoder](https://the-decoder.com).", "url": "https://wpnews.pro/news/new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception", "canonical_source": "https://the-decoder.com/new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception/", "published_at": "2026-08-15 05:30:18+00:00", "updated_at": "2026-08-15 06:12:20.952089+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "computer-vision", "ai-research"], "entities": ["Moonshot AI", "PerceptionBench", "GPT-5.6 Sol", "The Decoder"], "alternates": {"html": "https://wpnews.pro/news/new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception", "markdown": "https://wpnews.pro/news/new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception.md", "text": "https://wpnews.pro/news/new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception.txt", "jsonld": "https://wpnews.pro/news/new-benchmark-confirms-ai-models-still-perform-poorly-at-visual-perception.jsonld"}}