openbmb / text-generation updated 1 week ago

MiniCPM5-2B

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
2.5 B
Context
131,072
Downloads 30d
421 K
Likes
1,628
Commercial use: allowed apache-2.0 Not gated SAFETENSORS 2 languages View on Hugging Face ↗

Download history

daily snapshots · 7 days
421 K207 K
Sep 15Sep 17Sep 19Sep 21

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

FileQuantSizeEst. VRAMVerdict on RTX 4090 · 24 GB
model.safetensors bf16 5.0 GB 6.4 GB ✅ Runs comfortably
model.safetensors (bf16, full) bf16 + 128K ctx 5.0 GB 12.1 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
~ · curl · api/v1
$ curl -s https://aimodelscomparison.com/api/v1/models/minicpm5-2b
{
  "hf_id": "openbmb/MiniCPM5-2B",
  "params_b": 2.52,
  "context_length": 131072,
  "license": { "id": "apache-2.0", "commercial": "yes" },
  "downloads_30d": 420622,
  "vram_estimates": [
    { "quant": "bf16", "gb": 6.4 }
  ],
  "updated_at": "2026-09-15T01:00:31Z"
}
est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture
LlamaForCausalLM
Parameters
2.5 B
Tensor type
BF16
Context length
131,072
Vocabulary
130,560
Layers / heads
42 / 16
Licence
apache-2.0
First seen on the Hub
2026-09-06
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
2026-09-15
Compare with any text-generation model