Kimi VL A3B Thinking 2506
16.4B total, 3B active · 128k context · MIT · released 2025-06-21
visionreasoning
What to run it on in Germany
Fully in memory at 128k context, priced at retailers today.
Nothing we price holds this model fully in memory.
All hardware that runs it
Everything we price, best value first. Two cards split the model by layers.
Nothing we price can run it at this context.
Memory needed
Weights, KV cache and runtime buffers at each context length (KV cache at fp16).
| Quant | File |
|---|
Run it
Copy a command for the runtime you use.
No published builds to run yet.
Notes
Long-decode thinking model that generates up to 32K tokens; vLLM 0.9.1 with flash-attn explicitly installed is recommended to avoid CUDA out-of-memory. Requires trust_remote_code and supports images up to 3.2 million pixels plus video input.
Questions
- What licence is Kimi VL A3B Thinking 2506 released under?
- MIT, which allows commercial use.