🧩 Model Card: MiniCPM-V-4.7-1B

• Type: Image-Text-to-Text
• Think: Toggleable
• Tool Calling Support: Yes
• Base Model: openbmb/MiniCPM-V-4.7-1B
• Quantization: Q4_K
• Max Context Length: 256k tokens
• Default Context Length: 32k tokens (change default)
• Set Context Length at Launch

▶️ Run with FastFlowLM in PowerShell:

flm run minicpm-v-4.7:1b

📝 Note

• Images are processed the same way as in the original MiniCPM-V: the model sees a downscaled view of the whole image, and large images are also split into up to 9 tiles. Because of this, the --img-pre-resize (-r) option has no effect on this model.
• Image TTFT stops growing at about 1080p: from there up, every image is split into the maximum 9 tiles, so a 4K image costs about the same as a 1080p one. Each tile is about 448 × 448 pixels, so very fine detail in large images is downscaled.
• Video and audio input are not supported.