RESOURCES

Blog

Ideas, updates, and practical guidance from the Snorkel team.

Image for Data 2.0 and the research era of AI data

Data 2.0 and the research era of AI data

Announcing Snorkel’s $350M fundraise to build the frontier lab for AI data.

September 22, 2026
All articles
Sort: Newest
Paper2Agent: Turning Research Papers Into Interactive AI Agents
NEW
Paper2Agent: Turning Research Papers Into Interactive AI Agents

Jiacheng Miao (Stanford HAI) presents Paper2Agent, a framework that turns a research paper and its code into an interactive AI agent, scoring 91.2% on 300 benchmark questions against 80.3% for Claude Code with the repo.

Nov 19, 2026 •
Learn more about Paper2Agent: Turning Research Papers Into Interactive AI Agents
Training agents for professional work when success is rare
NEW
Training agents for professional work when success is rare

We trained Qwen3.8-27B on 1,105 professional tasks from Snorkel’s GDPval+ series. After about 102K training rollouts, the model reaches 40.7% pass@1 on OpenAI’s GDPval gold set, up from 31.9%. Its deliverables now beat those of human experts in 44.0% of LLM-judged head-to-head comparisons, compared with 36.4% for the base model. The gains come from two ingredients working together, expert-built training…

Nov 19, 2026 •
Learn more about Training agents for professional work when success is rare
Fable 5.2 vs Fable 5.1: Coding Benchmark Results
NEW
Fable 5.2 vs Fable 5.1: Coding Benchmark Results

Fable 5.2 pass@1 score lands at 71% compared to Fable 5.1 score of 53% on a set of 26 frontier coding tasks. The mid-run terminal failures that appeared in the analysis of Fable 5.1 nearly disappear when examining Fable 5.2.

Nov 12, 2026 •
Learn more about Fable 5.2 vs Fable 5.1: Coding Benchmark Results
Terminal-Bench 4.1 hardens the verifiers behind the scores
NEW
Terminal-Bench 4.1 hardens the verifiers behind the scores

Terminal-Bench 4.1 is now available. Snorkel AI contributes to Terminal-Bench as a task author, a data partner, and a supporter through the Open Benchmarks Grants program. The release keeps the same 66 tasks, with verifier fixes, reproducibility improvements, hardening against reward hacking, and a new 63-task CPU-only dataset. Terminal-Bench and real terminal work Terminal-Bench measures whether an AI agent can…

Nov 09, 2026 •
Learn more about Terminal-Bench 4.1 hardens the verifiers behind the scores
Intelligence per Watt: Measuring Local AI Efficiency
Intelligence per Watt: Measuring Local AI Efficiency

Jon Saad-Falcon and Avanika Narayan (Stanford, Hazy Research) present Intelligence per Watt, a metric that divides task accuracy by power draw to compare local AI models and hardware, and show it improved 5.3x from 2023 to 2025.

Nov 05, 2026 •
Learn more about Intelligence per Watt: Measuring Local AI Efficiency
Ping or Peek: When agents prefer to peek at rewards over asking users for help
Ping or Peek: When agents prefer to peek at rewards over asking users for help

The field of AI continues to suffer from two ongoing trends that we demonstrate here may be directly linked: (1) agents do not consistently ask users for clarification when tasks are ambiguous, and (2) agents sometimes inspect accessible reward-function code to infer the answer expected by a benchmark. Our hypothesis: pressure towards autonomous task completion incentivizes agents to bypass users…

Nov 05, 2026 •
Learn more about Ping or Peek: When agents prefer to peek at rewards over asking users for help
Joules-to-GDP: Measuring the economic value of AI per unit of energy
Joules-to-GDP: Measuring the economic value of AI per unit of energy

The model that produces the most valuable work isn’t always the best use of your energy. This is one of the interesting findings in Joules-to-GDP, a paper released today by researchers at Stanford, UC Berkeley and UT Austin. Of the five models they tested, GLM 5.2 delivered the most net value per task, but Gemma 4 31B delivered 56% of…

Learn more about Joules-to-GDP: Measuring the economic value of AI per unit of energy
Checking the Rubrics: RIFT Results on Five Public Benchmarks
Checking the Rubrics: RIFT Results on Five Public Benchmarks

Rubric-based evaluation is now a common way to score the outputs of language models and agents (Hashemi et al., 2024; Wei et al., 2025), and rubrics also serve as the source of reward signals in training (Gunjal et al., 2025; Liu et al., 2026; Ding, 2026). In both uses the rubric is the measurement instrument, and the validity of any…

Learn more about Checking the Rubrics: RIFT Results on Five Public Benchmarks
Frontier Data Summit 2026 Recap
Frontier Data Summit 2026 Recap

As agents get more capable, the data and evaluations that measure them have to become a science in their own right. Here are the sessions, takeaways and videos from the Frontier Data Summit.

Oct 16, 2026 •
Learn more about Frontier Data Summit 2026 Recap
1 2 … 41
Image

Join our newsletter

For expert advice, the latest research, and exclusive events.
By submitting this form, I acknowledge I will receive email updates from Snorkel AI, and I agree to the Terms of Use and acknowledge that my information will be used in accordance with the Privacy Policy.