MiniCPM5-2B-GGUF
We are releasing MiniCPM5-2B, the second model in the MiniCPM5 series, following MiniCPM5-1B. It is a dense 2B Transformer that scales up the same training recipe, built for on-device, local deployment, and resource-constrained scenarios, reaching 2B-class open-source SOTA.
Params
2.0 B
Context
—
Downloads 30d
192 K
Likes
384
Download history
daily snapshots · 33 days
▲ 20 K in the last 30 days (9.6%)
291 K172 K
Sep 26Oct 6Oct 16Oct 25
291 K172 K
Sep 23Oct 4Oct 15Oct 25
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| MiniCPM5-2B-Q4_K_M.gguf | Q4_K_M | 1.6 GB | 2.5 GB | ✅ Runs comfortably |
| MiniCPM5-2B-Q8_0.gguf | Q8_0 | 2.7 GB | 3.7 GB | ✅ Runs comfortably |
| MiniCPM5-2B-F16.gguf | GGUF | 5.0 GB | 6.3 GB | ✅ Runs comfortably |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ ollama run minicpm5-2b-gguf # pin the quantization explicitly $ ollama run minicpm5-2b-gguf-q4_k_m
$ huggingface-cli download openbmb/MiniCPM5-2B-GGUF-GGUF \
MiniCPM5-2B-Q4_K_M.gguf --local-dir .
$ llama-cli -m MiniCPM5-2B-Q4_K_M.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/minicpm5-2b-gguf
{
"hf_id": "openbmb/MiniCPM5-2B-GGUF",
"params_b": 2.00,
"context_length": null,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 191713,
"vram_estimates": [
{ "quant": "GGUF", "gb": 6.3 },
{ "quant": "Q4_K_M", "gb": 2.5 }
],
"updated_at": "2026-09-23T01:00:43Z"
}
Specifications
- Parameters
- 2.0 B
- Licence
- apache-2.0
- First seen on the Hub
- 2026-09-05
- Training datasets
- openbmb/Ultra-FineWeb, openbmb/UltraX-Preview, openbmb/Ultra-FineWeb-L3, openbmb/UltraData-Math, openbmb/UltraData-Code, openbmb/UltraData-SFT-2605, openbmb/UltraData-SFT-Agent-2609, openbmb/UltraData-RL-2609
- Added to our catalog
- 2026-09-23
Compare with any text-generation model