What are the hardware requirements for Agents A1 4B?

Agents A1 4B requires approximately 9.1 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Agents A1 4B and how large is it?

Agents A1 4B is developed by InternScience, a model with 4b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Agents A1 4B?

Agents A1 4B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Agents A1 4B by InternScience — VRAM 9.1GB

Name: Agents A1 4B
Author: InternScience

Agents A1 4B is an open-source language model by InternScience. Features: 4b LLM, VRAM: 9.1GB, License: apache-2.0, LLM Explorer Score: 0.37.

Arxiv:2606.30616 Agentic Conversational Endpoints compatible Image-text-to-text Qwen3 5 Region:us Safetensors Sharded Tensorflow Vision Vlm

Model Card on HF 🤗: https://huggingface.co/InternScience/Agents-A1-4B

Agents A1 4B Parameters and Internals

LLM Name	Agents A1 4B
Repository 🤗	https://huggingface.co/InternScience/Agents-A1-4B
Model Size	4b
Required VRAM	9.1 GB
Updated	2026-08-09
Maintainer	InternScience
Model Type	qwen3_5
Model Files	5.4 GB: 0-of-2 3.7 GB: 1-of-2
Model Architecture	Qwen3_5ForConditionalGeneration
License	apache-2.0
Model Max Length	262144
Transformers Version	4.57.0.dev0
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Errors	replace

Rank the Agents A1 4B capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Agents A1 4B

Best Alternatives	Context / RAM	Downloads	Likes
AgentHorse 4B	256K / 8.4 GB	973	7
Qwen3.5 4B	0K / 9.3 GB	6575502	801
NuExtract3	0K / 0.2 GB	441570	324
...en3.5 4B 3D Coder MTP Repacked	0K / 9.3 GB	2132	8
LightOnOCR 3 4B	0K / 9.1 GB	2395	71
FrogNano 4B 2609	0K / 9.3 GB	733	113
RuBio 4B	0K / 9.3 GB	3881	0
LightOnOCR 3 4B MTP Repacked	0K / 9.3 GB	662	1
Qwen3.5 4B Base	0K / 9.3 GB	227814	79
UserRepGen 4B	0K / 9.1 GB	1008	13

Note: green Score (e.g. "73.2") means that the model is better than InternScience/Agents-A1-4B.