Is Ling-3.2-flash state of the art?Pending benchmarks.Announced Nov 5, 2026; no independent results yet.
Ant Ling's native multimodal MoE model in the Ling series: 560B total parameters with about 25B activated per token, a 1M-token context window, and native image, text and video input. Positioned for complex visual understanding, multimodal long-horizon agents, 3D understanding and interaction, and medical multimodal work. Available through the Ant Ling API (OpenAI-compatible model id Ling-3.2-flash) and on Computrix demos. Open weights were not on inclusionAI Hugging Face at tracking time.
The details so far
- Lab
- Ant Ling (InclusionAI, Ant Group)
- Announced
- Nov 5, 2026
- API
- Live · Ant Ling developer platform; OpenAI-compatible model id Ling-3.2-flash
- Open weights
- Closed · Not published on inclusionAI Hugging Face as of 2026-11-05 (Ling-3.1-flash and Ling-3.1-flash-1019 are; this checkpoint is not)
- Parameters
- 560B total / 25B active
- Architecture
- Mixture-of-Experts; native multimodal (vision fused in the architecture, not an external vision module)
- Context
- Up to 1M tokens
- Modalities
- Image, text and video input
- Focus areas
- Complex visual understanding, multimodal long-horizon Agent, 3D understanding and interaction, medical multimodal
- Demo
- Computrix (Ant Group) per the lab's launch materials
- Related
- Separate from Ling-3.1-flash, Ling-3.1-flash-1019 and Ling-3.1-flash-base
Tags marked unsure are not stated on the lab's own pages.
What would let us call it
Verdicts use only independent boards we already track. Two consecutive snapshots are required before any crown.
- Publicly announced with a source
- Usable by the public (API or weights)
- Listed on an independent board we track
- Enough results for a rank in a category
- Holds across two snapshots (crowns only)
What the lab claims
Self-reported. Shown for context; not used in any verdict.
Lab claim: on OmniDocBench1.5, 92.21 vs 90.56 for Opus 5 (max).
Lab claim: on ZEROBench (Sub), 43.71 vs 39.52 for Opus 5 (max).
Lab claim: on Vision2Web, 85.86 vs 80.20 for Opus 5 (max).
Sources
LABLing model docs (Ant Ling developer platform)developer.ant-ling.com/en/docs/models/ling/ · Nov 5, 2026LABAntLingAGI on X: Ling-3.2-flashx.com/AntLingAGI/status/2118428246664761379 · Nov 5, 2026Tracking continues until an independent board lists it.