Ling 3.0 Flash VL

The Artificial Analysis article positions Ling-3.0-flash-VL as a comparator while covering the finance-focused Ling-3.0-flash-Fin release. Ling-3.0-flash-VL scored 25 on the Artificial Analysis Intelligence Index and shares a 256K token context window with the Fin sibling built on the same MoE base. On the Finance & Accounting Index, Ling-3.0-flash-Fin matched Ling-3.0-flash-VL at 24, with the Fin variant achieving higher business knowledge accuracy (17% vs. 11%) but also higher hallucination (33% vs. 19%). Ling-3.0-flash-VL scored 1225 Elo on GDPval-AA v2 and 986 on AA-Briefcase for professional agent tasks, and 16% on AutomationBench-AA.

Ant Group's inclusionAI lab released Ling-3.0-flash-VL on 10 September 2026 as an open-weights vision-language model under the MIT licence. It is a mixture-of-experts design with 124 billion total parameters, activating 5.5 billion per token, and ships alongside fp8, fp4, and int4 quantizations. The model accepts text, images, and video inputs and holds a 262,144-token context window, making it the multimodal counterpart to the earlier text-only Ling-3.0-flash. The architecture combines a ViT encoder with a small two-layer projector and VideoRoPE positional encoding, enabling timestamp-aware queries over long video. Its 42-layer backbone alternates Kimi Delta Attention and Gated MLA at a five-to-one ratio, keeping long-context memory costs manageable. Thinking mode is on by default but can be disabled, and inclusionAI positions the release as strengthening language reasoning while adding visual perception and visual-agent capabilities over its text predecessor.

The Baidu Baike entry corroborates the September 9, 2026 release of Ling-3.0-flash-VL by Ant Group as the Bailing series' first native multimodal model. BF16 and FP8 weights were open-sourced on Hugging Face and ModelScope, with a free trial available on Ling Studio, and a 256K token context window. The architecture uses an arbitrary-resolution visual encoder, VideoRoPE, and a 42-layer hybrid language backbone alternating KDA and Gated MLA at a 5:1 ratio. The visual feedback closed-loop mechanism continuously corrects outputs in medical, frontend, and GUI automation tasks. On Artificial Analysis Intelligence Index v4.1.1, it improved by 4 points over the text-only version.

Ant Group's open-source team released and open-sourced Ling-3.0-flash-VL on September 9, 2026, as the first native multimodal large model in the Bailing series. The article from aibase confirms the model is built on the MoE architecture of Ling-3.0-flash and natively supports image, text, and video inputs with a 256K token context window. Ling-3.0-flash-VL carries 124B total parameters with 5.5B activated per inference, inheriting Ling-3.0-flash's Agent-execution efficiency. It introduces a visual feedback closed-loop mechanism ("Observe → Act → Verify → Correct") for scenarios such as medical report interpretation, frontend code generation, and GUI automation, significantly improving result reliability.