HTTP 200 verified twice
Finding your next stop…
Loading the latest directory information.
Loading the latest directory information.
Agents-A1-4B is a 4B-parameter dense transformer agentic model from InternScience (Shanghai AI Lab) with a 262K context window. It supports function/tool calling, agentic reasoning, and strong performance on BrowseComp, GAIA, and related agent benchmarks.
Start here for the decision-making essentials: what Agents A1 4B does, who it is for, how it is accessed, and the first-party sources behind this profile.
See official pricing
Read-only public captures of Agents A1 4B’s homepage. Screenshots are dated, never live embeds, and open full-screen.
Published specifications and use cases for Agents A1 4B. Repository dates are kept separate from the model's original release date.
These are publisher-reported results. Test conditions and benchmark versions can differ.
Short answers to the questions buyers and builders commonly ask about Agents A1 4B. Each answer cites the shared ledger below, where every source is listed once.
Long-horizon search, engineering, scientific research, instruction following, and tool-cal
Agents‑A1 is a long-horizon agentic model from InternScience, built to scale heterogeneous agentic abilities across multiple domains including Long‑horizon Search, Engineering, Scientific Research, Instruction Following, and Tool-calling.
Compatible with Hugging Face Transformers, vLLM, and SGLang serving frameworks
These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, etc.
2026.7.14
2026.7.14: 🔥🔥 The 4B model has been released.
262K (262144 tokens)
--context-length 262144
4B
We release the dense model Agents-A1-4B with only 4B parameters
Text (text-only mode available)
Text-Only (skips vision encoder to free KV cache memory)
This profile connects the jobs Agents A1 4B is described as handling with its delivery model, access options and the subjects used to match it to related products in this directory.
Each fact points to a recorded source, making it easy to distinguish verified product information from claims that need checking.
HTTP 200 verified twice
InternScience/Agents-A1-4B · Hugging Face
Model family: Agents-A1 · Model version: Agents-A1-4B · Parameter count: 4B
Agents-A1
Agents-A1-4B
4B
2026.7.14
262K (262144 tokens)
Dense transformer
Text (text-only mode available)
English and Chinese
Open-source model weights publicly available on Hugging Face
A concise view of the jobs, capabilities and integrations described in the recorded product sources.
Documented product formats, platforms and official distribution destinations. Availability can vary by region and plan.
A concise history of software releases and material product changes. Events appear only when a dated source supports what changed.
Showing the newest updates and meaningful milestones. Open an entry for its summary and source.
The public model repository was created on Hugging Face. This repository date may differ from the model's original release date.
View source [1]Facts, answers, structured details, milestones and primary resource links cite this shared ledger. Each external page appears once; release tags from the same GitHub project are grouped under one release history.
These tools share a workflow, capability or audience with Agents A1 4B. They may complement it rather than replace it, and are not presented as integrations or endorsements.
Natively supports function calling and tool integration with APIs, code interpreters, search engines
Decomposes complex tasks into sub-steps, plans ahead, and adapts strategy based on intermediate results
Compatible with Hugging Face Transformers, vLLM, and SGLang serving frameworks
InternScience (developed by InternAgent Team at Shanghai Artificial Intelligence Laboratory)
Long-horizon search, engineering, scientific research, instruction following, and tool-calling agentic tasks
Quantized model variants available in the Agents-A1 collection (including mlx-community versions)
BrowseComp: 66.8; XBench-DS-2510: 90.0; GAIA: 95.1; FrontierScience-Research: 33.3; IFEval: 94.8
Trained using a domain-grounded knowledge-action infrastructure that jointly constructs actions, observations, and verifier outcomes
Dense model with only 4B parameters · Released on 2026-07-14 · Developed by InternAgent Team at Shanghai Artificial Intelligence Laboratory
4B
4B parameters (dense)
2026-07-14
Natively supports function calling and tool integration with APIs, code interpreters, search engines, and other external tools
Agentic reasoning: decomposes complex tasks into executable sub-steps, plans ahead, and adapts strategy based on intermediate results
Long-horizon Search, Engineering, Scientific Research, Instruction Following, and Tool-calling agentic workloads
Open-source model weights and configuration files released on Hugging Face
Developed by InternAgent Team at Shanghai Artificial Intelligence Laboratory (InternScience)
Series of quantized model variants released (with additional MLX-community quantizations for Mac)
Model name: Agents-A1-4B · Parameter count: 4B (dense) · Release date: 2026-07-14
Agents-A1-4B
Agentic reasoning — decomposing complex tasks into executable sub-steps, planning ahead, and adapting strategy based on intermediate results
262144
Hugging Face Transformers, vLLM, SGLang
Single GPU with 262K context window using vLLM or SGLang
OpenAI-compatible API served at http://localhost:8000/v1
Open source — model weights, evaluation code, and technical report publicly released
Reported benchmark scores: BrowseComp 66.8, XBench-DS-2510 90.0, GAIA 95.1, FrontierScience-Research 33.3, IFEval 94.8
