SPARKITSPARKIT 101: from pip install to a cited research report
An end-to-end walkthrough using a real HLE-Gold question: install the SDK, send the call, read the report, and see the agent trace behind a 54.4% benchmark answer.
Read post →← Back to blog
SPARKIT is live
By SPARKIT Research Team · Updated July 14, 2026
Updated July 14, 2026: Benchmark figures in this launch post now match the current evaluation. The canonical methods, exact comparators, and limitations live on the benchmark pages.
Today SPARKIT goes live: a scientific research agent in an API.
One call deploys an agent that retrieves and synthesizes the relevant literature, performs any analyses the question requires, and returns a Markdown report. On HLE-Gold, the gold-standard subset of Humanity's Last Exam, SPARKIT scores 54.4%, compared with 39.0% for direct GPT-5.6-Sol and 34.9% for direct Claude Opus 4.8.
The shape of the API is intentionally tiny:
from sparkit_science import SparkitClient
client = SparkitClient(api_key="sk_sparkit_...")
report = client.research(
"What is the role of BRCA1 in homologous recombination?"
)
print(report)
Get an API key at app.sparkit.science/signup. Pricing is on /pricing, start with Try-it for $10 and 5 queries.
We'll post deeper writeups, benchmark updates, and engineering notes here. Star the GitHub org for release notifications.
More from the blog
Hallucinated vs. fetched: a GAIA case study on a verifiable question
We ran a single GAIA question through SPARKIT, direct Claude Opus 4.7, and direct GPT-5.5. SPARKIT fetched Nature's archive, counted 1,002 articles, and answered correctly. Both direct LLMs invented different article counts and confidently landed on the wrong answer.
Read post →How to think about the deep-research tool landscape
Perplexity, ChatGPT Deep Research, Gemini Deep Research, Claude Research, Elicit, SPARKIT — the deep-research category has converged on the basics. The remaining differences are about audience, deployment surface, and how each tool treats citations. A map of who builds what for whom, and where SPARKIT fits.
Read post →