SAGE Celer 2.6 模型强化逻辑验证与南亚语言支持，提升推理准确性。
SAGE Celer 2.6 Technical Card
- 通过逆向推理管道训练，减少复杂任务中的错误传播与幻觉。
- 在数学、编程和通用智能基准上表现优异，延迟低。
- 原生支持南亚语言，对天城文字符有定制分词器，兼顾英語能力。
我们推出 SAGE Celer 2.6，是 SAGEA 系列通用型 Celer 模型的最新版本，提供 5B、10B 及 27B 参数量级。该模型经过大规模架构改进，并在未公开的数据集上进一步预训练。采用 SAGEA 自研的逆向推理（IR）流程，使 Celer 2.6 能自主验证逻辑路径，有效降低复杂推理任务中的级联错误与幻觉。模型还集成端到端视觉编码器，实现原生多模态功能，避免适配器方法常见问题。在数学、编程与通用智能基准（ACUMEN）测试中表现极具竞争力，且延迟极低。最重要的是，Celer 2.6 特别优化了南亚语言支持，配备针对天城文脚本的定制分词器，在尼泊尔语和印地语上表现良好，同时不牺牲英语推理能力。
原文摘要 · Abstract (English)
We introduce SAGE Celer 2.6, the latest in our line of general-purpose Celer models from SAGEA. Celer 2.6 is available in 5B, 10B, and 27B parameter sizes and benefits from extensive architectural modifications and further pre-training on an undisclosed model. Using our Inverse Reasoning (IR) pipeline, SAGEA natively trains Celer 2.6 to validate its own logic paths, minimizing cascading error and hallucination in complex reasoning tasks. Celer 2.6 also boasts natively integrated multimodal functionality with an end-to-end vision encoder to avoid common pitfalls in adapter-based approaches. Celer 2.6 provides highly competitive results on mathematics, coding, and general intelligence benchmarks (ACUMEN), along with low latency. Most importantly, Celer 2.6 is specifically optimized for South Asian language support, with a custom tokenizer for the Devanagari script and strong performance in both Nepali and Hindi without sacrificing English reasoning ability.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。