Terminal-Bench-Science: Evaluating AI agents on scientific research workflows (terminal-bench-science.ai)
55 points by matt_d 5 hours ago
boorang 41 minutes ago
I got pretty good mileage out of context engineering, adding my personal coding heuristics to my AGENTS.md and referencing subdocuments on a "when doing X, consult Y" pattern. I assume others are doing similar things, but I was pretty surprised when I was able to get it to generate code that is pretty close to what I would do if I was doing it manually. I'm curious if scientists and mathematicians are doing things like that. "When I see X, I typically immediately check Y" or whatever their domain heuristics look like.
respectattentio 24 minutes ago
Like you were reading my mind. I was waiting for such benchmark to land. This will improve models for such scientific research workflows.
AI should have started with science from the beginning, not after 4 years.
I am building on top of it with agents to improve scientific workflows.
johnnyApplePRNG 2 hours ago
Not surprised to see Claude significantly higher in scientific intelligence than Sol.
You can tell that Claude really does grasp a wide array of highly specific scientific and mathematical nuances... where's codex is just basically for coding and that's it.
That's the feel I get from the both of them anyways and I've used both on the 20x plan for the past week at length.
Don't get me wrong, codex is great at finding bugs and building games. It's great.
jerpint 2 hours ago
The fact that opus 5 is outperforming fable is odd to me
From personal experience, opus 5 feels net inferior to fable on almost every aspect (for coding tasks)
ianjbutler 18 minutes ago
IMHO it's about task depth (fable) vs breadth (opus). Fable is great at tracing and debugging sometimes, but otherwise shorthand for confabulation. It's persistent but ungovernable, struggles to switch contexts, and goes insane with with too much freedom to explore. Don't point it at anything that looks like a "system" for actual work (but mapping or planning might be ok). Opus is maybe not as creative, but it's more stable and more trustworthy. Opus driving Fable could be awesome, but Fable unleashed/unsupervised on longer horizon tasks or things that require more methodical changes on lots of components seems like a disaster every time I try it.
Since this kind of thing is always down to harness, project-type, and other structural constraints, of course your mileage may vary. Fable is probably great for pen-testing, or as a decision-making kernel of other kinds of applications, and way better than Opus at those things. Probably fine for code-review or changing a codebase of a few thousand lines in any language. Actually building that codebase or changing an even bigger one? Woof.
How this fits in with science? IDK but I bet other existing causal reasoning benchmarks might tell the whole story and this is back to stability again. Sometimes having a smart idea is really important! But more often it's important to just not forget what you were doing. What was I talking about? Oh look a squirrel
jhbadger 2 hours ago
This is measuring on scientific tasks though. I haven't used Fable lately but when I was playing with it when it was new its "safety" features made it practically impossible to use for biomedical science -- it once refused to work on a pipeline of mine that was analyzing pathogenicity islands in bacteria (presumably because it had guardrails to that effect to stop potential bioterrorists and the like)
akshay_akula 4 hours ago
Evals on actual research workflows is the right direction, most agent benches are toy tasks.
a2ff6eeb0 an hour ago
I wonder how long it's going to be before self improvement encompasses hardware and materials science, not just code. It's exciting, soon we'll be able to fully hand off scientific, mathematical, and technical progress over to the machines, and then we can fully lay back.
rubslopes 4 hours ago
I'm glad to see that GPT Sol beats Opus at least in Mathematical Sciences, because that's my need right now, and I much prefer GPT's prose style.
mlmonkey 3 hours ago
Sad to see no mention of Gemini ...
vatsachak 3 hours ago
Damn. These things aren't AGI... but I don't care.
Luna is good enough for me to give a parser spec and have it write one.
daveguy an hour ago
How is an llm parser_spec to parser better than something like lex?
vatsachak 36 minutes ago
Because sometimes parsers can have weird inputs, like structured Excel files