Ministral 3 3B 2512
12B
- Context
- 8Ktokens
- Released
- 2y ago
- Size
- 2Bparameters
Small open Gemma 2 instruction model for lightweight chat and on-device use
How it compares
Size class
Tiny2BAn open-weight model under 5 billion parameters, small enough for phones, laptops and edge devices.
Open models of a similar size from other labs
Vs. your stack
Add the models you use to see how this one compares.
Sign inAdditional source evidence
- Source model
- google/gemma-2-2b-it
- Source family
- gemma
- Last updated
- 2024-07-31
- Model context
- 8K
- Model output limit
- 8K
- Open weights
- Yes
- File attachments
- No
- Reasoning
- No
- Tool calling
- No
- Temperature control
- Yes
Downloadable weights
Complete source record (JSON)
{
"id": "google/gemma-2-2b-it",
"name": "Gemma 2 2B IT",
"description": "Small open Gemma 2 instruction model for lightweight chat and on-device use",
"family": "gemma",
"last_updated": "2024-07-31",
"attachment": false,
"reasoning": false,
"tool_call": false,
"temperature": true,
"modalities": {
"input": [
"text"
],
"output": [
"text"
]
},
"limit": {
"context": 8192,
"output": 8192
},
"release_date": "2024-07-31",
"open_weights": true,
"license": "Gemma Terms of Use",
"weights": [
{
"label": "Hugging Face",
"url": "https://huggingface.co/google/gemma-2-2b-it"
}
]
}Text provenance
Gemini app/webGoogle SynthID text watermarking
Google DeepMind says text generated in the Gemini app and web experience is watermarked with SynthID by adjusting token probabilities.
Official explanationFrequently asked questions
Running locally
Gemma 2 2B IT needs an estimated 1.9 GB at Q4, 2.9 GB at Q8 and 5.1 GB at FP16. All figures are estimates covering weights plus about 15% and 0.5 GB for a short context.
Yes. A 24 GB, 16 GB, 12 GB or 8 GB GPU all fit the model at FP16, with an estimated 5.1 GB of VRAM.
Yes. At Q4 the estimated VRAM is 1.9 GB, and assuming macOS lets the GPU use about 75% of unified memory, the smallest common Mac memory size that fits is 8 GB.
Privacy and data
Yes, Gemma 2 2B IT has open weights, so you can self-host it for full control of your data.
Capabilities and specs
The context window is 8K tokens (8,192) and the maximum output is 8K tokens (8,192).
No. Gemma 2 2B IT accepts text input only.
No, tool calling (function calling) is not supported.
No, a reasoning or thinking mode is not supported.
Quality and benchmarks
Gemma 2 2B IT holds rank 352 on Arena Text, which is Arena's own rank, with a score of 1200.
Matched on model kind and release date, with no comparable price, the closest alternatives are BGE Multilingual Gemma2 by BAAI, DCLM 7B by Apple and OLMoE 1B-7B by Ai2.
BGE Multilingual Gemma2Gemma 2 2B IT vs BGE Multilingual Gemma2DCLM 7BGemma 2 2B IT vs DCLM 7BOLMoE 1B-7BGemma 2 2B IT vs OLMoE 1B-7B
Versions and lineage
Both have an 8K token (8,192) context. Gemma 2 2B IT was released on July 31, 2024 and Gemma 2 27B on June 27, 2024, and Gemma 2 27B costs $0.65 input and $0.65 output per 1M tokens.
Licensing and access
Gemma 2 2B IT is open weights, not open source, under the Gemma Terms of Use license.
The weights can be downloaded from the Hugging Face repository google/gemma-2-2b-it.
Basics
Gemma 2 2B IT was made by Google DeepMind and released on July 31, 2024.
Gemma 2 2B IT has open weights.
Generated from Vertical Tabs data, updated November 11, 2026.
Discussion
No comments yet
Be the first to share your thoughts.