LFM-40B

Sarvam-30B is an open source, next-generation large language model designed as a unified system for both real-time conversational AI and deep reasoning workloads, built with a strong focus on multilingual intelligence and practical deployment. The 30B model is optimized for speed and efficiency, using a Mixture-of-Experts (MoE) architecture that activates only a subset of parameters per request, enabling high throughput, low latency, and deployment even in resource-constrained environments such as local machines or edge systems. It delivers strong performance in conversational tasks, coding, and reasoning while achieving state-of-the-art results across more than 20 Indian languages, making it highly effective for multilingual applications and voice-based systems. It represents a dual-tier architecture, a fast, deployable “conversational workhorse”, leveraging MoE designs to reduce compute cost while maintaining high performance.

Qwen3.5-35B-A3B is part of the Qwen3.5 “Medium” model series, designed as a highly efficient, multimodal foundation model that balances strong reasoning ability with practical deployment requirements. It uses a Mixture-of-Experts (MoE) architecture with 35 billion total parameters but activates only about 3 billion per token, allowing it to deliver performance comparable to much larger models while significantly reducing computational cost. The model integrates a hybrid attention mechanism that combines linear attention with standard attention layers, enabling efficient long-context processing and improved scalability for complex tasks. As a native vision-language model, it can process both text and visual inputs, supporting use cases such as multimodal reasoning, coding, and agent-based workflows. It is designed to function as a general-purpose “AI agent,” capable of planning, tool use, and structured problem solving rather than just conversational responses.