01
Scientific task requirements
AI for science
Terminal-Bench-Science
Scientific computing tasks for calibration, experimental design and uncertainty analysis.
Benchmark / task domain
Terminal-Bench format
The work
Turn scientific models and research inputs into executable tasks.
Agents work with scientific models and research inputs to estimate parameters, compare explanations and assess predictions. Required outputs include numerical results and a clear account of the method used.
Delivery can include
- 01
- Harbor-format task packages with Markdown instructions and task.toml configuration
- 02
- Docker environment definitions, research inputs and output schemas
- 03
- Reference solutions and numerical evaluation code
How we evaluate the task
Define how to evaluate the result.
02
Numerical results and interpretation
03
Completeness of the analysis
WORK WITH ASTRAFLUX
What should your
model be able to do?
Share the capability, task types and delivery requirements. We’ll discuss the right program and representative tasks.
Contact us norman383@astraflux.spaceBenchmark names indicate reference directions. Task scope and delivery specifications are agreed for each project.