Vendor · Language models · Image models
inclusionAI has 15 models in the directory: 13 language models and 2 image models.
In the directory: 13 language models and 2 image models. 15 of them with open weights that can run locally.
Every model there ever was. Models with a profile lead to every benchmark, price and source.
| Ling 3.0 Flash | Aug 2, 2026 | Active | below 1 | $0.021 / $0.063 | 243 | 262K | 39 |
| Ring 2.6 1T | May 14, 2026 | Active | n/a | $0.30 / $2.50 | 118 | 262K | 35 |
| Ming-flash-omni-2.0 | Feb 11, 2026 | Open source | n/a | n/a | n/a | n/a | 42 |
| Ling 3.1 Flash | n/a | Active | n/a | n/a | 221 | 1M | 41 |
| Ling 3.1 Flash 1019 | n/a | Active | n/a | n/a | n/a | n/a | 35 |
| Ling 3.1 Flash Base | n/a | Active | n/a | n/a | n/a | 262K | 29 |
| Ling-3.0-flash-Fin | n/a | Open weights | n/a | n/a | n/a | n/a | below 1 |
| Ling-3.0-flash-VL | n/a | Open weights | n/a | n/a | n/a | n/a | below 1 |
| Ling-3.0-tiny | n/a | Open source | n/a | n/a | n/a | 262K | 45 |
| LLaDA 2.2 Mini | n/a | Open weights | n/a | n/a | n/a | n/a | 39 |
| ZwZ2-27B | n/a | Active | n/a | n/a | n/a | 262K | 30 |
| ZwZ2-4B | n/a | Active | n/a | n/a | n/a | 262K | 29 |
| ZwZ2-9B | n/a | Active | n/a | n/a | n/a | 262K | 30 |
| Ming-Image-0.1-Design | n/a | Open weights | n/a | 34 |
| Ming-Image-0.1-Design-Layer | n/a | Open weights | n/a | 29 |
Language model · Aug 2, 2026
Ling 3.0 Flash stands out for low active compute alongside a large model capacity. It has 124 billion total parameters but activates about 5.1 billion per token; 35 KDA, 7 MLA layers, 512 experts, and top-8 routing support a 262,144-token context. InclusionAI presents it as the next generation after Ring 2.6 1T.
Language model · May 14, 2026
Ring 2.6 1T combines one trillion total parameters with about 63 billion active per token. The open model brings a 262,144-token context and high or extra-high reasoning intensity together.
Language model · Feb 11, 2026
Ming-flash-omni-2.0 combines image editing, video conversations, and controllable audio output in one model. InclusionAI’s omnimodal checkpoint accepts text, images, video, and audio, and generates text, images, and audio.
Language model · n/a
Ling 3.1 Flash activates 25 billion of its 560 billion parameters per token. InclusionAI’s hybrid reasoning model handles 262K context tokens on Vercel AI Gateway.
Language model · n/a
Ling 3.1 Flash 1019 is inclusionAI's upgraded checkpoint in the Ling 3.1 Flash series. It is kept separate from the existing Ling 3.1 Flash profile and from the base checkpoint. The model card lists 560 billion total parameters, 25 billion active parameters per token, and a hybrid MoE architecture.
Language model · n/a
Ling 3.1 Flash Base is the base checkpoint in the Ling 3.1 Flash series. inclusionAI recommends it for further training, adaptation, and research, not as a ready-to-use chat assistant. The model card lists 560 billion total parameters and 25 billion active parameters per token. The released weights are BF16 and come from FP8 training followed by Warmup-Stable-and-Merge checkpoint merging.
Language model · n/a
Ling-3.0-flash-Fin is InclusionAI’s first finance-specialized model in the Ling series, according to the company. It builds on Ling 3.0 Flash with continued training on financial data. Its focus includes reconciling reporting periods and figures, valuation models, and spreadsheet formulas.
Language model · n/a
Ling-3.0-flash-VL adds native image and video input to Ling 3.0 Flash. VideoRoPE represents spatial positions and temporal order, supporting questions about events in longer videos. InclusionAI adds a visual encoder to the language backbone for this purpose.
Language model · n/a
Ling-3.0-tiny brings InclusionAI’s hybrid attention architecture to a 7.9B checkpoint with 1.3 billion active parameters. Three Kimi Delta Attention layers alternate with one Multi-Head Latent Attention layer. The provider documents local tests on DGX Spark and Apple Silicon Macs.
Language model · n/a
LLaDA 2.2 Mini uses Levenshtein editing to modify sequences during parallel generation through DELETE and INSERT control tokens. The MoE diffusion model has 16 billion parameters excluding embeddings, activates 1.4 billion, and supports a provider-reported 128K context.
Image model · n/a
Ming-Image-0.1-Design is a 6-billion-parameter model for interfaces, infographics, posters, and other text-rich designs. It produces complete compositions, returns RGBA images with transparent backgrounds, and recommends 2048 x 2048, with 1024 x 1024 for faster runs.
Image model · n/a
Ming-Image-0.1-Design-Layer takes a flattened design and a layer plan, then decomposes it into requested RGBA layers, unlike Ming-Image-0.1-Design. The recommended working resolution is 1024, or 512 for faster runs.
3 more current models are in the model list above.
| Vendor | Models | Model types | Best language model | Top 10 | Open weights | New in 2026 |
|---|---|---|---|---|---|---|
| Anthropic | 33 | 33 language models | Claude Opus 5.5 (Rank 1 to 3) | 4 | 0 | 13 |
| OpenAI | 63 | 46 language models, 4 image models, 2 video models, 9 audio models, 2 specialist models | GPT-6 Astra (Rank 4 to 8) | 2 | 4 | 27 |
| 66 | 43 language models, 4 image models, 6 video models, 11 audio models, 2 specialist models | Gemini 4 Carbon (Rank 1 to 5) | 2 | 19 | 26 | |
| Meta | 32 | 29 language models, 1 image model, 1 video model, 1 specialist model | Muse Wonder Full (Rank 6 to 9) | 2 | 26 | 7 |
| DeepSeek | 23 | 23 language models | DeepSeek V4.1 Flash 1017 (Rank 13 to 20) | 0 | 23 | 9 |
| Moonshot AI | 12 | 12 language models | Kimi K3.1 Pro (Rank 10 to 13) | 0 | 10 | 7 |
| xAI | 16 | 12 language models, 2 image models, 2 video models | Grok 4.7 (Rank 18 to 24) | 0 | 1 | 12 |
| inclusionAI | 15 | 13 language models, 2 image models | not ranked yet | 0 | 15 | 3 |
By release date. Grey: retired models.
3 releases since January 2026, by release date.
Prices per 1M tokens for the current language models, according to inclusionAI or the provider linked in each model profile.
| Current model | Input | Output | Context |
|---|---|---|---|
| Ling 3.0 Flash | $0.021 | $0.063 | 262K |
| Ring 2.6 1T | $0.30 | $2.50 | 262K |
The directory lists 15 inclusionAI models: 13 language models and 2 image models. It counts every model inclusionAI ever released, including retired ones. 15 are currently available, the newest being Ling 3.0 Flash, Ring 2.6 1T and Ming-flash-omni-2.0.
Ling 3.0 Flash, released on Aug 2, 2026. The timeline below shows every release by date.
Through the API per 1M tokens (input / output): Ling 3.0 Flash $0.021 / $0.063. The prices of all 2 current language models are in the price table.
Yes, 15 of the 15 models have open weights. Each model profile lists how much memory it needs.
Ling 3.0 Flash with 243 tokens per second according to Artificial Analysis. A typical model manages 98.
The largest is Ling 3.1 Flash with 1 M tokens (≈ 1,500 A4 pages). Every model's value is in the model list.
Data as of Nov 1, 2026