Model Detail

Ling-3.0-flash-VL-fp4 is a code generation model with 32.2B parameters released by inclusionAI. The model is registered under the image-text-to-text pipeline tag on Hugging Face, distributed under the permissive mit license.

Ling-3.0-flash-VL-fp4 ships with 32.2B parameters. Total weight footprint is approximately 64.5 GB, which is the relevant figure when planning local-inference VRAM. The mit license is permissive, allowing commercial deployment and derivative work without per-seat fees, though attribution requirements still apply.

Ling-3.0-flash-VL-fp4 is best fit for code completion, repository-scale Q&A, and pair-programming integrations. It is a less obvious choice for one-shot generation of security-critical code without review. Treat this as a starting matrix rather than a benchmark verdict — the right deployment usually depends on the specific evaluation suite that mirrors your workload.