{"version":1,"type":"story","url":"https://digestai.news/story/longcat-debuts-longcat-2-5-with-1-6t-parameters-and-visual-token-suppo","json":"https://digestai.news/story/longcat-debuts-longcat-2-5-with-1-6t-parameters-and-visual-token-suppo.json","markdown":"https://digestai.news/story/longcat-debuts-longcat-2-5-with-1-6t-parameters-and-visual-token-suppo.md","slug":"longcat-debuts-longcat-2-5-with-1-6t-parameters-and-visual-token-suppo","headline":"LongCat debuts LongCat-2.5 with 1.6T parameters and visual token support","summary":"LongCat released LongCat-2.5 on November 1, 2026, a multimodal mixture-of-experts model with 1.6 trillion total parameters and 48 billion activated per token. The model improves on LongCat-2.0 with native visual token processing, enabling unified handling of text and images. It uses discrete visual tokens, Residual Vector Quantization, and a shared semantic space for visual and textual inputs, with offline preprocessing to optimize performance and memory efficiency.\n\nLongCat-2.5 incorporates long-horizon reinforcement learning for complex coding tasks, including full-project builds and agentic coding. Training includes synthesized tasks, verifiable tests, and real-world failure trajectories. The model achieves 86% higher per-device throughput in long-sequence training and 43.3% higher prefill throughput in inference compared to baselines. Benchmark results show 55.4 on 3DCodeBench and 78.6 on Vision2Web, with API access available via OpenAI- and Anthropic-compatible endpoints. Pricing starts at $0.30 per million uncached input tokens, with discounts for cached inputs.","keyPoints":["LongCat-2.5 has 1.6 trillion parameters and 48 billion activated per token, with native visual token support for multimodal tasks","Model achieves 86% higher per-device throughput in training and 43.3% higher prefill throughput in inference compared to baselines","API access available with 1M-token context window and pricing starting at $0.30 per million uncached input tokens"],"whyItMatters":"LongCat-2.5’s multimodal capabilities and long-horizon reasoning could advance agentic AI for complex coding, design, and project tasks, while its efficiency improvements may lower deployment costs for enterprises.","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"entities":{"companies":["LongCat"],"models":["LongCat-2.5","LongCat-2.0"],"people":[]},"firstPublishedAt":"2026-11-02T21:42:04Z","updatedAt":"2026-11-02T21:42:04Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"Unite.AI","title":"LongCat-2.5 Debuts With 1.6T Parameters and Native Visual Tokens","url":"https://unite.ai/longcat-2-5-debuts-with-1-6t-parameters-and-native-visual-tokens","publishedAt":"2026-11-02T21:42:04Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"LongCat debuts LongCat-2.5 with 1.6T parameters and visual token support\", 2 November 2026, https://digestai.news/story/longcat-debuts-longcat-2-5-with-1-6t-parameters-and-visual-token-suppo","publisher":"Digest AI","title":"LongCat debuts LongCat-2.5 with 1.6T parameters and visual token support","datePublished":"2026-11-02T21:42:04Z","url":"https://digestai.news/story/longcat-debuts-longcat-2-5-with-1-6t-parameters-and-visual-token-suppo"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}