Model library / inclusionAI
Ling-3.0-flash is a 127.5 B parameter hybrid-linear language model from inclusionAI for reasoning and long-context applications.
Parameters in total
Context length in tokens, as published
Weights at the released precision
Licence declared on the repository
Overview
Ling-3.0-flash is a native hybrid-linear model with 124 B total and 5.1 B active parameters, built by inclusionAI. The publisher describes it for agentic workflows, coding, general knowledge, mathematical reasoning and long-context tasks, emphasising speed, efficiency and production deployment. It incorporates alternating KDA and MLA attention layers and a sparse MoE design.
In practice the model has 127.5 B total parameters, activates 5.1 B per token, and supports a 262,144-token context window. It is released under the MIT licence, stored in bfloat16 format, and uses the BailingMoeV3ForCausalLM architecture without expert layers. Deployment targets 4×141 GB GPUs or equivalent high-memory nodes.
Specification
| Repository | inclusionAI/Ling-3.0-flash |
| Publisher | inclusionAI |
| Published | 2 August 2026 |
| Architecture | BailingMoeV3ForCausalLM |
| Model type | bailing_hybrid |
| Parameters | 127,486,405,600 |
| Layers | 42 |
| Hidden size | 2,560 |
| Attention heads | 32, 32 key/value heads |
| Context length | 262,144 tokens |
| Vocabulary | 157,184 tokens |
| Weight format | bfloat16 |
| Licence | mit |
Memory
Parameter count multiplied by the bytes each format uses. The key-value cache comes on top.
| Precision | Weights | Notes |
|---|---|---|
| bf16 | 237.5 GiB | Full precision as released |
| fp8 | 118.7 GiB | 8-bit, near-lossless on most models |
| int4 | 59.4 GiB | 4-bit, smallest footprint |
The cache costs 672 KiB per token at 16-bit, so the full 262,144-token context of one request needs about 168.0 GiB. Concurrency multiplies that number, not the weights.
That figure is an upper bound: this model uses compressed key-value attention, so the real cache is a fraction of the figure above.
Hardware
At the released precision, with 90% of the card given to the server and 2.5 GiB kept for the runtime. The context column is what is left for the cache on that setup.
| GPU | Cards needed | Free for cache | Context it holds |
|---|---|---|---|
| RTX 4090 24GB | more than 8 | — | — |
| RTX 5090 32GB | more than 8 | — | — |
| L40S 48GB | 8 | 88.1 GiB | about 137,528 tokens |
| A100 80GB | 4 | 40.5 GiB | about 63,254 tokens |
| H100 80GB | 4 | 40.5 GiB | about 63,254 tokens |
| RTX PRO 6000 96GB | 4 | 98.1 GiB | about 153,132 tokens |
| H200 141GB | 2 | 11.3 GiB | about 17,691 tokens |
Serving
A vLLM launch line for the shape above. Check the model card for a runtime the publisher recommends.
vllm serve inclusionAI/Ling-3.0-flash \ --tensor-parallel-size 8 \ --max-model-len 135168 \ --gpu-memory-utilization 0.90
Questions
The weights take about 237 GiB at the released precision. No single card in the table above holds that, so it needs several GPUs or a lower precision.
The repository declares mit. Read the licence text before commercial use: the name alone does not tell you what is allowed.
The configuration allows 262,144 tokens. Whether the whole window is usable depends on the memory left for the cache, which the hardware table works out per card.
This page is a reference, and Aiqre does not serve this model today. We deploy open-weight models on dedicated EU hardware on request, so if you want an OpenAI-compatible endpoint for it, email [email protected] with the model name and roughly what volume you expect.
Also in the library
LLaDA2.2-mini is a 16.3 billion-parameter MoE diffusion language model with 128 k token context, supporting Levenshtein editing.
Ling-3.0-flash-Fin is a 127.5 billion-parameter finance-enhanced language model for long-context financial research and analysis.
Ling-3.0-tiny is a 7.9 billion-parameter hybrid MoE language model with 1.3 billion activated parameters per token.
Laguna S 2.1 is a 118 billion-parameter mixture-of-experts model for agentic coding with long-context support.
Hosting
Tell us the model and roughly what volume you expect. We reply with a price and an OpenAI-compatible endpoint, with your prompts never stored.