inclusionAI
Ant Group's open-source AI lab behind the Ling, Ring and Ming model families.
| Model | Released | Weights | Hosts | Cheapest host | Input | Output |
|---|---|---|---|---|---|---|
| Ling 3.1 Flash−75% | Sep 2026 | Open · MIT | 2 | DeepInfra | $0.075 | $0.23 |
| Ling 3.0 Flash Sante (inclusionAI) | Sep 2026 | — | 1 | Novita AI | $0.075 | $0.22 |
| Ling 3.0 Flash VL (inclusionAI) | Sep 2026 | Open · MIT | 2 | DeepInfra | $0.06 | $0.18 |
| Ling 3.0 Flash Fin (inclusionAI) | Sep 2026 | Open · MIT | 1 | Novita AI | $0.075 | $0.22 |
| Ling 3.0 Flash | Jul 2026 | Open · MIT | 2 | Novita AI | $0.06 | $0.18 |
5 of 5 models · prices per million tokens at the cheapest host, read Oct 23, 09:51 UTC
Deals now
- −75%Ling 3.1 Flashat DeepInfra
About inclusionAI
- Founders
- Ant Group (corporate founder; no individual named)2
- First model
- Ling-Lite and Ling-Plus, Mar 2025 (earliest documented)6
- Official API operated from
- China (PRC)5
- Mission
- "Provide a Sustainable and Responsible Intelligence Option for the Learners and Builders"2
inclusionAI was founded by Ant Group, and its GitHub organisation describes its projects as open-source work from Ant Group.1, 2
Its stated mission is to provide a sustainable and responsible intelligence option for learners and builders.2
In a technical report submitted to arXiv on 7 March 2025, the Ling team presented Ling-Lite, with 16.8 billion parameters, and Ling-Plus, with 290 billion, both mixture-of-experts models.6
The Ring reasoning models are derived from Ling, and the Ming multimodal model Ming-lite-omni is derived from Ling-lite.8, 9
In October 2025 the Ant Ling API platform launched with Ling-1T, and Ring-1T followed on 14 October.10
The platform changelog lists Ling-2.6-1T in April 2026 and Ling-3.0-flash in July 2026.10
The Ant Ling site now lists three model families: Ling, Ring and Ming.3
Besides language models, inclusionAI publishes the AReaL reinforcement-learning framework and the AWorld multi-agent platform.11
The models are published with open weights under MIT or Apache licences.1, 7
The Ling service is provided by a company in Hangzhou, and its privacy policy says personal information collected in mainland China is stored there.4, 5
Official links
- Website inclusion-ai.org
- API docs developer.ant-ling.com/en/docs
- Pricing developer.ant-ling.com/en/docs/models/price
- GitHub github.com/inclusionAI
- Hugging Face huggingface.co/inclusionAI
- Blog inclusion-ai.org/blog
- X x.com/ant_oss
- X x.com/AntLingAGI
- LinkedIn linkedin.com/company/ant-ling
Sources
- inclusionAI · GitHub · github.com · read Oct 23
- INCLUSION AI · inclusion-ai.org · read Oct 23
- Ant Ling · ant-ling.com · read Oct 23
- Ant Group - Home · antgroup.com · read Oct 23
- 百灵大模型服务隐私权政策 · render.alipay.com · read Oct 23
- [2503.05139] Every FLOP Counts: Scaling a 300B Mixture-of-Experts LING LLM without Premium GPUs · arxiv.org · read Oct 23
- inclusionAI/Ling-1T · Hugging Face · huggingface.co · read Oct 23
- Ring: A Reasoning MoE LLM Provided and Open-sourced by inclusionAI | INCLUSION AI · inclusion-ai.org · read Oct 23
- Ming-Omni: A Unified Multimodal Model for Perception and Generation | INCLUSION AI · inclusion-ai.org · read Oct 23
- Changelog · developer.ant-ling.com · read Oct 23
- About | INCLUSION AI · inclusion-ai.org · read Oct 23
Every fact above links to its source and the date it was read; last checked Oct 23. Not affiliated with inclusionAI. Corrections or removal on request: contact.