Trust and security · 2026-09-22
Ling 3.0 Flash VL: the free period is ending, planning the vision workload move
Covers that inclusionai/ling-3.0-flash-vl's free offer ends on 22 September 2026, that no free card with image and video input replaces it, and how to move a vision workflow to a paid model.
The model page states this date in its own description
inclusionai/ling-3.0-flash-vl's catalog description directly states that the free offer ends on 22 September 2026, and that no free card carrying image and video input replaces it. This isn't a removal or an error on the provider side; it's been listed in the catalog as a time-limited free offer from the start.
If you picked this model for a prototype or a low-priority flow because it was free, note that after this date you won't find the same image+video input combination on a free card, and you'll need to move to a paid vision model.
Three things to check before migrating
First, verify the target model supports both image AND video input together; some vision models accept image only and don't support video input. Second, compare reasoning behavior: on this model thinking is on by default and the only way to turn it off is sending `reasoning_effort: "none"` (intermediate levels have no effect on this model) — retest the target model's effort levels with your own task set. Third, check prompt-cache behavior; since this was a free card, the cache's gain stayed in latency only, while on a paid model the same cache now also lowers direct cost.
If your flow needs audio input or schema-conforming structured output, this model lacked both; separately verify whether the target model closes those gaps.
- Verify the target model supports image + video input together.
- Retest reasoning_effort behavior on the new model with your own tasks.
- This model lacked audio input and schema-conforming JSON output; check the target.
Don't leave the migration until the date arrives
An integration still sending requests to this identifier once the free period ends will run into failing requests after that date. Treating the migration as a planned maintenance window, rather than a production outage, avoids a disruption to user experience.
Frequently asked questions
What happens if I send requests to this identifier after 22 September 2026?
The model page states the free offer ends on this date; check the current model page before sending requests for the exact error behavior, since the row may be fully removed on the provider side or moved to a paid tier.
Are this model's reasoning_effort levels the same as other Ling models?
No, the only level with effect here is 'none' (fully turns thinking off); some other Ling models may carry intermediate levels like minimal/low/medium/high — verify this separately for each model.