InclusionAI / inclusionai/ling-3.0-flash-vl

Ling 3.0 Flash VL - access through LLMTR

Ling 3.0 Flash VL is InclusionAI's vision build of Ling 3.0 Flash, activating roughly 5.1 billion of its 124 billion parameters per token. Alongside text it accepts image and video input, so it is used for screenshot and document reading, visual question answering, and clips where the order of what happens matters. It offers a 256K context window and produces at most 32,768 tokens per response. It reasons by default, returning that chain separately in `reasoning_content`, whose tokens are counted inside `completion_tokens`. To turn reasoning off, send `reasoning_effort: "none"` or append the `:none` suffix to the model id - that is the only level with a measured effect on this model, so the others are not offered even though the upstream accepts them. Native tool calling works in every reasoning state, including `tool_choice` set to `required` or a named function. Repeated prefixes are served from a prompt cache; since this card is free the benefit is LATENCY, not a lower price. It does not accept audio input and cannot produce schema-enforced JSON output. The free period ends on 22 September 2026: after that date this card retires and requests return an error explaining where to go next, and nothing is charged until you change the model id yourself. Note that no free card with image and video input replaces it, so plan for a metered vision model if you depend on those.