Ai2's open AstaBrief 8B writes cited research reports 3.5 times faster than Asta's Claude mode
Ai2 open-sourced AstaBrief 8B, which writes cited research reports in 51.1 seconds against 178.5 for Asta's Claude-powered mode. Labs that cannot send unpublished research questions to a hosted model can now run a cited-report generator on their own servers.
The Engineer · Build desk
What happened
- AstaBrief already runs as Fast mode in Asta's Generate a report feature, and Ai2 published the training data along with the weights.
- Ai2 started from Qwen3-8B and put most of its effort into post-training data, evaluation and the report-generation scaffolding.
- Training drew on tens of thousands of real research queries, filtered with a focus on citations and paired with preference data.
- Ai2 also released an example workflow that researchers can adapt to build reports from their own PDFs on local machines.
- Most training and evaluation finished in 2025, and Ai2 says it has not rerun the full evaluation against today's frontier models.
Why it matters
- What it costs cost A lab that wants to retrain or extend AstaBrief can use the published data with SFT and DPO and skip the judge-in-the-loop RL setup that Ai2 says can be unstable and expensive.
- Who has to decide decision Teams choosing between AstaBrief and a hosted model need their own citation-grounding comparison against current models, because Ai2's parity test stopped at the 2025 frontier.
- What it limits constraint The 51.1-second figure belongs to Asta's pipeline, so a lab running the PDF workflow on its own GPUs changes both retrieval and hardware and has to time the model again.
Ai2 gives two speed figures, and they measure different things. The first, in the post's words, is "nearly an order-of-magnitude reduction in report generation time compared to the proprietary models we tracked" [16]. The second covers the whole Asta pipeline. There, Fast mode averages 51.1 seconds per report against 178.5 for the Claude-powered Thinking mode, so Fast is about 3.5 times faster [3]. Each Fast report arrives 127.4 seconds sooner [15].
Some of the gap between roughly 10x and 3.5x is presumably time spent outside the model, in pipeline stages that a smaller model does not shorten. Some may come from the comparison itself. The larger figure is measured against the proprietary models Ai2 tracked [16], and the smaller one against a mode Asta ships today [2].
The design change I would credit first is in the scaffolding. Ai2 rebuilt report generation to write the full report in one pass instead of section by section [12]. A sectioned pipeline makes several model calls per report, each with its own prompt. One pass makes a single long call over the research question and the retrieved excerpts [1]. The 3.5x figure compares two shipped modes. It does not show how much of the gain came from the scaffolding and how much from the smaller model [3].
On training, Ai2 considered reinforcement learning. Its DR Tulu work showed RL can improve long-form reports from open-weights models, especially with judge models in the training loop [6]. Ai2 chose supervised fine-tuning and DPO instead [6], and I think it chose correctly. "RL-based training can be unstable and expensive," Ai2 wrote [7].
The quality bar covered answer quality, relevance, structure and citation grounding [14]. Grounding matters in Asta because users come back to generated reports later and treat them as working research artifacts [17]. Ai2 framed the project as a test of whether a small open model could match the proprietary models it was using on report quality while cutting generation time and serving costs [13]. Ai2 is careful about what the test shows. Its results, Ai2 wrote, "are best read as evidence about the particular training and system design choices we tested" [9].
What to watch
- A rerun of Ai2's full evaluation against current frontier models would show whether quality parity holds beyond the 2025 comparison set.
- Independent timings of AstaBrief 8B on self-hosted GPUs with the PDF workflow would show how much of the 3.5x gain survives outside Asta's pipeline.
- An ablation separating the one-pass scaffolding from the model swap would show which change bought the speed.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence55
- Adoption18
- Hype gap+22
- Incentives55
- Confidence60
Perspective Coverage
3 publishers- Builder
- Builder 55%
- Operator
- Operator 33%
- Investor
- Investor 12%
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
AstaBrief 8B is a model that turns a research question and retrieved literature excerpts into a cited report.
- [2]
AstaBrief is available in Asta's Generate a report feature as Fast mode alongside Claude-powered Thinking mode, and Ai2 is open-sourcing the model and its training data so others can study, reproduce and build on the approach.
- [3]
Across the full Asta pipeline, Fast mode averages 51.1 seconds per report compared with 178.5 seconds for Thinking mode, about 3.5x faster.
- [4]
Developing AstaBrief required tens of thousands of real research queries, citation-focused filtering and preference data.
- [5]
Ai2 started from Qwen3-8B and focused most of its effort on the post-training data, evaluation and surrounding report-generation scaffolding.
- [6]
Ai2's DR Tulu work showed reinforcement-learning-based methods can improve long-form report generation for open-weights models, especially with judge models in the training loop; Ai2 considered that path for AstaBrief but focused on a simpler recipe built around supervised fine-tuning and direct preference optimization.
- [7]
RL-based training can be unstable and expensive.
- [8]
Most of the training and evaluation was completed in 2025, so the proprietary models used to generate training data and as comparison points reflect the frontier at that time; Ai2 has not rerun the full evaluation against today's frontier models.
- [9]
are best read as evidence about the particular training and system design choices we tested
- [10]
Open weights let institutions run AstaBrief on their own infrastructure, which Ai2 says is necessary when research questions reveal sensitive or unpublished work.
- [11]
Ai2 is releasing an example workflow that researchers can adapt to create reports from their own PDFs, as a starting point for local report generation.
- [12]
Ai2 redesigned the report-generation pipeline to write the full report in one pass rather than section by section.
- [13]
Ai2 tested whether a small, open model trained specifically for scientific report generation could match the report quality of the proprietary models it was using, while reducing generation time and serving costs.
- [14]
Ai2's goal was an open-weights model with answer quality, relevance, structure and citation grounding for long-form scientific synthesis.
- [15]
A Fast-mode report finishes 127.4 seconds sooner than a Thinking-mode report across the full Asta pipeline.
- [16]
nearly an order-of-magnitude reduction in report generation time compared to the proprietary models we tracked
- [17]
Many Asta users return to generated reports later, treating them as working research artifacts.
Get this ledger inside ChatGPT. Ask about any of our stories and the answer brings its figure, each claim’s verdict and the sources. Free, no account.
Sources
3 independent publishers whose own reporting we read for this story.
- dev.toAstaBrief 8B: How AllenAI Trained a Small Open Model to Generate Cited Scientific Reports
1 article · October 2, 2026
- huggingface.coOpen-sourcing AstaBrief, the fast report-generation model in Asta
1 article · October 2, 2026
- runtimewire.comAi2 releases an open 8B model for cited science reports
1 article · October 2, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.