moonshotai/Kimi-VL-A3B-Thinking
moonshotai/Kimi-VL-A3B-Thinking is a moonshotai record on this register. The Hugging Face model API lists the pipeline tag as image-text-to-text. The safetensors index gives an exact parameter count of 16,407,657,776. Architectural context length (max_position_embeddings) is 131,072. The listed licence is mit. The Hub repository was created on 9 Apr 2025, which is the release date this register uses.
Primary source: the Hugging Face model card for moonshotai/Kimi-VL-A3B-Thinking. Fields below come from that card, the Hub model API, and the repository’s own config.json when those files exist. A field we cannot source is omitted rather than guessed.
| Organisation | moonshotai |
|---|---|
| ReleasedRepo creation date on Hugging Face | 9 Apr 2025 |
| Last modified | 30 Jan 2026 |
| ParametersExact, from the safetensors index | 16.4B 16,407,657,776 |
| Context lengthmax_position_embeddings — architectural maximum, not necessarily served | 128K 131,072 |
| Licence | mit |
| Pipeline / taskpipeline_tag on the Hugging Face model API | image-text-to-text |
| File formats | safetensors |
| Architecture | kimi_vl |
| Layers | 27 |
| Vocabulary | 163,840 |
| DownloadsAt last check | 16.3k |
| Likes | 454 |
| Gated | No |
How to run
Safetensors weights are on the Hugging Face repository. The parameter count on this page came from that index.
Load it from huggingface.co/moonshotai/Kimi-VL-A3B-Thinking. This page does not add install steps that the Hub files do not state.
Agent capabilities
Derived from the model’s own chat template, with the match counts behind each finding.
- ○Tool callingno evidence in template
- ○Reasoning / thinkingno evidence in template
- ●System prompt2 matches in template
- ●Image input4 matches in template
- ○Video inputno evidence in template
- ○Audio inputno evidence in template
Questions this record can answer
- When was Kimi-VL-A3B-Thinking released?
- The Hugging Face repository was created on 9 Apr 2025, which is the date this register uses.
- How many parameters does Kimi-VL-A3B-Thinking have?
- The safetensors index gives 16,407,657,776 parameters, shown here as 16.4B. That figure is exact, not rounded.
- What is the context length of Kimi-VL-A3B-Thinking?
- 131,072 tokens. That is max_position_embeddings from the model's own config.json — the architectural maximum, not necessarily the context a provider serves.
- What licence is Kimi-VL-A3B-Thinking under?
- The Hub card lists mit.
- Is Kimi-VL-A3B-Thinking gated?
- No. The Hugging Face model API does not mark this repository as gated.
- Does the Kimi-VL-A3B-Thinking chat template show tool calling?
- The chat template was read and no tool-calling evidence was found.
Where each field came from
- Hugging Face model API — releasedAt, lastModified, downloads, likes, task, license, gated, parameters. Checked 29 Sept 2026.
- config.json — contextLength, layers, vocabSize, architecture. Checked 29 Sept 2026.
- chat template — capabilities. Checked 29 Sept 2026.
First recorded 24 Sept 2026, last checked 29 Sept 2026. This record is compiled automatically from the sources above. Where a field is blank, we could not source it and have not guessed. Canonical repository: huggingface.co/moonshotai/Kimi-VL-A3B-Thinking.