Chota added new models for different business scenarios

The same agent can answer a client, read a long conversation, work with documents, help a manager qualify a request, or run a more complex tool-based workflow. For Chota, the right model is not always the largest one: it is the model that fits the task, latency target, and token budget.

Simple client conversations usually need speed and predictable cost. Complex integrations, long instructions, analytics, and multi-step workflows need stronger planning, longer context, and better tool use.

DeepSeek V4 Flash is useful for fast, cost-efficient customer replies, FAQ handling, first-step qualification, and high-volume conversations where responsiveness matters.

DeepSeek V4 Pro is better for long instructions, complex logic, knowledge-base analysis, integrations, and cases where the agent needs to hold a large context.

Kimi K2.6 is useful for agentic work, coding, long task chains, and workflows where the model has to stay stable across multiple tool-driven steps.

Claude Sonnet 4.6 fits complex wording, careful reasoning, large-document analysis, high-quality writing, and scenarios where answer reliability matters more than raw speed.

Gemini 3.1 Flash Lite Preview is a good candidate for high-frequency workloads: quick translation, classification, moderation, extraction, and lightweight multimodal processing.

MiMo-V2.5-Pro is a strong option for heavy agent workflows, complex software/workflow reasoning, and long-horizon tasks with many tool calls.

MiMo-V2.5 is useful when a business needs a balance of multimodality, agent capability, and cost: images, documents, video fragments, and long customer requests.

Mistral Medium 3.5 is an open multimodal model for agentic and coding use cases, useful when control, portability, and document workflows matter.

Clients do not need to choose from model names manually. During setup, we map the model to the real business task: cheap fast replies where they are enough, stronger reasoning where the workflow is sensitive, and multimodal or long-context routing where the agent needs it.

This helps control cost without lowering quality in the important parts of the workflow: inquiries, booking, CRM handoff, and non-standard customer questions.