:wave: i was on the team! AMA.

some headlines

- 3000 rubrics on code quality. First benchmark to measure: "would this code get actually merged?"

- 20+ expert open-source maintainer created tasks on their own repos to capture their opinion & taste.

- total 1000+ hours of real life software maintainer work captured in dataset. ON TOP of that, 40+ hours of real human work to turn that real life work into well validated and structured tasks with rubrics (even more work to turn tasks/prompts from devin-infra-specific to pluggable coding agent)

- results in 81% lower false positive rate than SWE-Bench Pro

- High quality bar: many QA stages & each task manually reviewed by Cognition researchers (examples in post)

Opus 4.8 scores 13% on FrontierCode Diamond.

one of my goals was also to datamine interesting stuff even on the easy tasks. for example, if you squint you can see the answer to "WTF Happened in late 2025" with coding models: https://x.com/swyx/status/2064081945567580323

▲

tedsanders 9 hours ago | parent | next [-]

Very cool! So glad to see people building and sharing evals that are better than SWE bench.

I'm curious - any particular reason you didn't put error bars on the graphs? Seems like it could be helpful when there are only 50 unique problems in the diamond set.

▲

swyx 8 hours ago | parent [-]

*50 unique problems but 20-40 rubrics per problem (something I had to keep reminding people internally who were unimpressed with the N)

simple answer is our reporting was pass@5. feel like you'd need like 50+ runs to have reasonable confidence intervals, which somehow i dont see other people do, so i also didnt insist on it.

hoping to work with <prominent third party evals shop> to get this on their infra and evaluated along with whatever the industry standard is.

	▲	tedsanders 7 hours ago \| parent [-]
		Makes sense, thanks. I suppose error bars are tricky if trying to handle problem-to-problem variance, rubric-to-rubric variance, and run-to-run variance all at once.

▲

glerk 6 hours ago | parent | prev | next [-]

This looks really great, more thoughtful than any benchmark that I've seen until now!

I'm curious if you're only interested in scoring frontier models or you would accept submission from custom harnesses? I am working on multi-model harnesses and would love to test them against your benchmark. Do you plan on releasing the tasks publicly?

	▲	swyx an hour ago \| parent [-]
		> Do you plan on releasing the tasks publicly? yep

▲

typs 8 hours ago | parent | prev | next [-]

What did you do around cross-harness testing? I don't see anything in the blog post about what harnesses were used in evaluation. SOTA benchmarks have consistently shown that frontier model performance is quite sensitive to what tools are exposed (e.g. str_replace vs. apply_patch) as the labs are RLing on their own harnesses. Did you do testing of the models in a standard setup or in their native harnesses?

▲

swyx 8 hours ago | parent [-]

yes well aware :) numbers shown are on "house" harnesses eg codex with gpt and claude code with opus.

fwiw we have examples of each model doing better on NON-house harnesses too - speaking jsut for myself i think the "the labs are RLing on their own harnesses" narrative is kinda overstated if you think through wanting to have any meaningful api business (often eg the labs will give guidance on what is prefered and the agent labs can easily match tool contract to that, which is to say, the "home turf advantage" isnt as large as you think it is if you try a little bit)

	▲	Bolwin 5 hours ago \| parent \| next [-]
		What is the "house" harness for minimax? They haven't released any
	▲	chris_st 7 hours ago \| parent \| prev [-]
		What "non-house" harnesses have you found to work best?

▲

fouc 2 hours ago | parent | prev | next [-]

I'm a bit disappointed that Opus 4.6 wasn't in this because the tokenizer changed quite a bit from 4.7 onward. I was so annoyed by 4.7 that I've been forcing 4.6 ever since. I've been annoyed by 4.8 a bit too, so I haven't felt the urge to move on.

▲

great_psy 9 hours ago | parent | prev [-]

How do you measure quality at scale ? Is there another model that determines if it adheres to codebase standard ?

	▲	swyx 9 hours ago \| parent [-]
		see Beyond Unit Tests and Novel Grading Methods in TFA. i think something like ~60% llm as judge rubrics and the rest as described. every rubric validated by maintainer. 3000 rubrics