Leaderboard / OpenBMB (MiniCPM)

MiniCPM-V 4.6 Thinking

Name: MiniCPM-V 4.6 Thinking
Author: OpenBMB (MiniCPM)

OpenBMB (MiniCPM) · MiniCPM-V · released 2026-05-08 · data checked 2026-11-02

MiniCPM-V 4.6 Thinking is a 1.3B-parameter dense model from OpenBMB (MiniCPM) (China), released 2026-05-08 under the Apache-2.0 license. It doesn't have enough reported benchmarks for an index score yet. It needs about 1.4 GB of memory at 4-bit with an 8K-token context.

Reasoning version of MiniCPM-V 4.6. The card reports only vision benchmarks (MMBench 84.1, MMMU 55.3).

Small LM Index

—

Needs 2 core benchmarks (has 0)

Parameters

1.3B

Dense

Price per 1M tokens

—

No hosted API on OpenRouter; self-host

Context window

262K

tokens

Benchmarks

Each bar shows this model's score. The tick marks the best tiny model on that benchmark.

No benchmark results for this release yet. Know of one? Send it.

Specifications

Lab: OpenBMB (MiniCPM) (CN · China)
Based on: Qwen3.5 0.8B (fine-tune or distillation)
Architecture: Dense transformer
Total parameters: 1.3B
Active parameters: All (dense)
Size class: Tiny
Context window: 262K tokens
Input modalities: text, image, video
Reasoning: Reasoning (thinking)
License: Apache-2.0
Openness: Open weights
Memory to run: ~1.4 GB at 4-bit, 8K context (more sizes)
Hugging Face: 43.9K downloads (30 days) · 33 likes

Weights on Hugging Face ↗

Compare with

MiniCPM-V 4.6 Thinking vs EXAONE 4.0 1.2B (1.3B, index 36.1)
MiniCPM-V 4.6 Thinking vs LFM2.5 1.2B Thinking (1.2B, index 30.1)
MiniCPM-V 4.6 Thinking vs LFM2.5 1.2B Instruct (1.2B, index 20.6)

Sources

https://huggingface.co/openbmb/MiniCPM-V-4.6-Thinking

Something wrong here? Report it.

Run it locally

Uses openbmb/MiniCPM-V-4.6-Thinking-gguf (Q4_K_M). Swap the tag for another size: F16, Q4_0, Q4_1, Q4_K_M, Q4_K_S, Q5_0, Q5_1, Q5_K_M….

Ollama

ollama run hf.co/openbmb/MiniCPM-V-4.6-Thinking-gguf:Q4_K_M

llama.cpp

llama-server -hf openbmb/MiniCPM-V-4.6-Thinking-gguf:Q4_K_M

Then open http://localhost:8080 for a chat window.

vLLM (GPU server)

vllm serve openbmb/MiniCPM-V-4.6-Thinking

Full-precision weights, for a server GPU.

Open in LM Studio

Memory needed: weights, context cache and runtime buffers.

Quantization	4K context	32K context	256K context
16-bit F16	3.2 GB	3.5 GB	6.3 GB
8-bit Q8_0	2.0 GB	2.3 GB	5.1 GB
4-bit Q4_K_M	1.4 GB	1.7 GB	4.5 GB

Check it against your GPU, Mac or phone →