Ling-3.0-tiny, a lightweight hybrid reasoning MoE model with 7.9B total parameters and only 1.3B activated parameters per token. It is designed to deliver strong reasoning and agentic capabilities at low inference cost, making advanced model capabilities.
detailed thinking off{{ end }}<|role_end|>{{ else }}<role>SYSTEM</role>{{ if .Think }}detailed thinking on{{ else }}detailed thinking off{{ end }}<|role_end|>{{ end }}
{{- range .Messages }}
{{- if eq .Role "user" }}<role>HUMAN</role>{{ .Content }}<|role_end|>{{ else if eq .Role "assistant" }}<role>ASSISTANT</role>
{{- if .Thinking }}
<think>{{ .Thinking }}</think>{{ else }}
<think></think>{{ end }}{{ .Content }}<|role_end|>{{ else if eq .Role "tool" }}<role>OBSERVATION</role>