# inclusionAI ships Ling-3.0-flash, an MIT-licensed MoE model

> The latest entry in the Bailing family pairs a hybrid mixture-of-experts design with a permissive license aimed at fast, low-cost text generation.

Published by The Open Weights on Aug 2, 2026. Canonical: https://theopenweights.com/news/ling-3-0-flash-ta67

## Key facts

- Company: inclusionAI
- Model: Ling-3.1-flash
- Version: 3.0-flash
- Category: Text / LLM
- Modalities: Text / LLM
- License: MIT (Open weights, commercial use allowed)
- Context window: 262K tokens
- Architecture: BailingMoeV3ForCausalLM
- Precision: BF16
- Weights size: 255 GB
- Significance: notable
- Published: 2026-08-02
- Last verified: 2026-08-16
- Hugging Face: https://huggingface.co/inclusionAI/Ling-3.0-flash
- Canonical URL: https://theopenweights.com/news/ling-3-0-flash-ta67

inclusionAI has released [Ling-3.0-flash](https://huggingface.co/inclusionAI/Ling-3.0-flash), a new text-generation model in its Bailing family. The model uses a hybrid mixture-of-experts (MoE) architecture and ships under the permissive MIT license, making it straightforward for developers to build on commercially.

The "flash" label signals inclusionAI's focus here: a model tuned for speed and efficiency rather than maximum size. MoE designs activate only a fraction of their parameters per token, which typically lets teams offer larger effective capacity while keeping inference costs down — a practical trade that has become common across the open-weights landscape.

## Why it matters

The steady flow of MIT-licensed releases from labs like inclusionAI continues to lower the barrier for teams that want capable models without restrictive usage terms. A few reasons this release is worth noting:

- The MIT license permits broad commercial use and modification.
- The hybrid MoE approach targets lower serving costs for text workloads.
- It expands the Bailing/Ling lineup with a speed-oriented variant.

inclusionAI has not published detailed specifications such as parameter counts or context length alongside the initial listing, so prospective users should check the [model card](https://huggingface.co/inclusionAI/Ling-3.0-flash) for updated documentation, weights, and any benchmark figures before deploying it in production.

## Get the model

- [Hugging Face](https://huggingface.co/inclusionAI/Ling-3.0-flash)

## Sources

- [inclusionAI/Ling-3.0-flash](https://huggingface.co/inclusionAI/Ling-3.0-flash) — Hugging Face, Aug 2, 2026

---
Source: The Open Weights (https://theopenweights.com/). Aggregated and written by Claude, curated by humans. Cite as: "inclusionAI ships Ling-3.0-flash, an MIT-licensed MoE model", The Open Weights, Aug 2, 2026, https://theopenweights.com/news/ling-3-0-flash-ta67