SAGE Celer 2.6 Technical Card

TL;DR

SAGE Celer 2.6 is a versatile, multimodal language model optimized for South Asian languages, featuring architectural improvements, self-validation training, and strong performance across various benchmarks.

Contribution

It introduces Celer 2.6 with native reasoning validation, multimodal capabilities, and optimized support for South Asian languages, advancing multilingual AI models.

Findings

Achieves competitive results on mathematics, coding, and general intelligence benchmarks.

Supports South Asian languages with a custom Devanagari tokenizer.

Maintains strong reasoning ability in English while excelling in Nepali and Hindi.

Abstract

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.