Model reference · open weights
LLaDA2.1-mini is an open-weight language model from inclusionAI, listed in the AxForge catalogue. AxForge can bring it up on EU-owned hardware for you on request — with the licence handled where one is required.
About
LLaDA2.1-mini 🚀 LLaDA2.1-flash is now live on ZenmuxAI! Try it via API 🛠️ or Chat 💬: https://zenmux.ai/inclusionai/llada2.1-flash LLaDA2.1-mini is a diffusion language model of the LLaDA series featuring the editing enhancement. It significantly improves inference speed while delivering strong task performance. Model Performance 🚀 Highlights + Error-Correcting Editable: Structural innovation of editable generation for dLLM + Speedy vs Quality Mode: The 16B mini model achieves ultra-fast inference under Speed Mode while remaining competitive across various tasks and under Quality Mode. + Reinforcement Learning on 100B-scale dLLM: Tailored algorithm and framework to enable reinforcement learning for large dLLM. 🗺️ What's Next + Powerful Agentic/Tool Use Capability with LLaDA: Next update will be equipped with powerful Agentic and long-distance tool-use capability. + Extreme Editing: Next update will feature stronger and more extensive editing capabilities, aimed at correcting more errors in parallel reasoning. + Explore More Training Paradigms: We want to explore more training paradigms than SFT and RL for dLLM. 📦 Model Variants 🔍 Model Overview LLaDA2.1-mini has the following specifications: + Type: Mixture-of-Experts (MoE) Diffusion Language Model + Total Parameters (Non-Embedding): 16B + Number of Layers: 20 + Attention Heads: 16 + Context Length: 32,768 tokens + Position Embedding: Rotary (RoPE) + Vocabulary Size: 157,184 🤗 Hugging Face Transformers Make sure you have transformers and its dependencies installed: Best Practices To achieve optimal performance, we recommend the following settings: 1. Sampling Parameters: We recommend the following general sampling parameters: blocklength=32, temperature=0.0, topp=None and topk=None. We are currently exploring more diverse sampling configurations. 2. Denoising Thresholds: There are three denoising params: threshold, editingthreshold and maxpoststeps. We recommend threshold=0.7, editingthreshold=0.5 for Quality Mode and threshold=0.5, editingthreshold=0.0 for Speed Mode. For both modes, we suggest setting maxpoststeps to a value greater than 5. We recommend 16 as a balanced default, which was used for most of our i
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | inclusionAI |
|---|---|
| Type | Language models |
| Parameters (lead) | 16.3B |
| Context | 32k tokens |
| Variants | 1 |
| Runs with | transformers |
| Released | 2026-02-09 |
| Popularity | 105k downloads / month |
| Likes | 124 |
| Licence | Open weights |
How it works
Variants
Open weights ship in several sizes and precisions. One page, all the variants — pick the one that fits your GPU. VRAM figures are estimates from model size.
| Variant | Params | Precision | VRAM | Fits 16 GB | Weights |
|---|---|---|---|---|---|
| LLaDA2.1-mini | 16.3B | BF16 | ~37.4 GB | ✓ | Weights ↗ |
Using it via the API
Once AxForge deploys llada2-1-mini for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (llada2-1-mini below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/chat/completions \
-H "Authorization: Bearer $AXFORGE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"llada2-1-mini","messages":[{"role":"user","content":"Hello"}]}'
Licence
Open weights under apache-2.0 — commercial use is permitted. Deploy it on AxForge EU hardware on request. Read the licence ↗