Skip to content
latestai.eulatestai.eu
AS

AstaBrief

by Allen Institute for AI

WebsiteGitHubHugging Face

Open 8B Qwen3-based model from Ai2 that turns a prompt and papers into a cited research report in one pass.

Available in EUSearch & ResearchLanguage ModelsModel

Updated EU status checked Last shown

AstaBrief at a glance

What is AstaBrief?
AstaBrief is a model from Allen Institute for AI. Open 8B Qwen3-based model from Ai2 that turns a prompt and papers into a cited research report in one pass.
Who makes AstaBrief?
AstaBrief is made by Allen Institute for AI.
Does AstaBrief work in the EU?
Available in the EU. Open weights are reported as downloadable from Hugging Face with no EU exclusion found; the Ai2 terms only reference US export controls and the privacy policy has EU/UK provisions. The specific license was not confirmed. Last verified 14 Nov 2026.
How much does AstaBrief cost?
Open source. Weights reported as openly downloadable; no pricing information found.
Where was AstaBrief shown?
In 1 creator video by AI Search. Most recently by AI Search on 4 Oct 2026, at 35:26 in “Gemini 4, GPT 6.1, Plasma conjecture, Napoleon code solved, OpenAI Dots, Claude Sonnet 5.5: AI NEWS”.

Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.

About

AstaBrief (also referred to as Astro Brief) is a research-report generation model from the Allen Institute for AI (Ai2), part of its Asta science-agent effort. According to a YouTube creator summary, it is an 8B-parameter model built on Qwen3 that takes a prompt together with a set of papers and writes a cited research report in a single pass.

The same source says it outperforms base Qwen3 and comes close to Dr. Tulu on this task, with both the model and its data released openly. The checkpoint is reported at about 32 GB. These details come from a secondary source and were not confirmed on the official pages provided, so the exact release date, license and benchmarks are unverified.

It is aimed at researchers and developers who want a self-hostable, inspectable model for literature synthesis, rather than a hosted service. Weights are presumably distributed through Ai2's Hugging Face organization.

Key features
  • 8B parameters, based on Qwen3
  • Generates cited research reports from a prompt plus supplied papers
  • Single-pass generation
  • Model and data reported as fully open
  • Reported to beat base Qwen3 and approach Dr. Tulu
1 moment

As seen on YouTube

Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.

Keep exploring

Similar AI tools

Closest neighbours by meaning, not just by category.

SA

Sage A1.5

Model · Alpha Innovator

1

35B mixture-of-experts open-weights agentic model for engineering, long-horizon search and scientific research.

Available in EULanguage Models
Release date n/aOpen source
A3

Apex 35B

Model · Apexin

1

35B-parameter model from Apexin, built on Qwen3.6-35B-A3B and tuned for automated research and robotics simulation.

UnverifiedLanguage Models
Release date n/a
AA

Agents A1

Model

1

35B Mixture-of-Experts open-weights model from InternScience for long-horizon agentic search, engineering and tool use.

Available in EULanguage Models
Release date n/aOpen source
Q3

Qwen 3.8 Flash Next

Model · Alibaba

2

Alibaba's Qwen3.8 Flash Next: a 125B-parameter MoE language model with about 6B active parameters, previewing Qwen 4 design.

Available in EULanguage Models
Release date n/aOpen source
Q3

Qwen 3.5 0.8B

Model · Alibaba

1

Small 0.8B-parameter open-weights language model in Alibaba's Qwen 3.5 family, aimed at lightweight and on-device use.

Available in EULanguage Models
Release date n/aOpen source
GA

GPT-6 Astra

Model · OpenAI

22

OpenAI's GPT-6 Astra, a frontier language model used for reasoning, agentic work and weight interpretation research.

UnverifiedLanguage Models
Release date n/aAPI