Cognition releases FrontierCode, a coding benchmark built by open-source maintainers to evaluate models on complex software maintenance
Opus 4.8-medium solved 32% of tasks at a 40x speedup
Positive users praise the FrontierCode benchmark for focusing on real mergeable code from OSS maintainers and see Opus 4.8 leading as a strong industry signal, while some negative users call the results fake or manipulated.
Most Activity
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
It's finally out!!! @METR_Evals found that more than half of SWEBench results is unmergeable slop. FrontierCode represents over 1000+ hours of maintainer validated software engineering work most frontier models cannot yet solve, much less solve with high quality.
Cog had IOI Gold medalists and top code maintainers Look At The Data — FrontierCode includes 3000+ rubrics covering code quality and anticheat reward hacking plaguing other benchmarks.
FC Diamond is so hard that Opus 4.8 scores 13.8%.
Three eras of AI coding : Three eras of benchmarks
2021 • Autocomplete : HumanEval 2023 • Passing Tests: SWEBench, TerminalBench 2026 • Maintainable Code: FrontierCode
to me the most beautiful chart when I requested a special historical run into all extant old models, the data was finding that the easiest third of FC tasks (in FC Extended) were rapidlly and suddenly solved over late 2025 - Opus almost doubled from a 41% pass rate to 74% in 4 months.
This describes the "WTF happened in Dec 2025" vibe shift that a lot of folks from @dhh to @karpathy have called out: it is the difference between getting 95% success in 2 rerolls vs 6, making it finally feasible to go up the next layer of abstraction in agentic coding, eg @GeoffreyHuntley's ralph loops or @bcherny's /goals or @steipete's "loops that prompt your agents" without fearing too much that things go off the rails.
My guess: as AI accelerates from here, each FrontierCode tier will saturate in sequence, hopefully ~annually. I've already asked the team to prepare FrontierCode 2027....
The old mountains will be destroyed. Their rubble becomes regolith. And from that regolith, the next model forest grows. Circle of life.
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
Opus 4.8 is the best coding model out there
FrontierCode by Cognition is probably the highest quality coding benchmark we have seen so far
it moves beyond just using unit-testing for scoring, it also tests for regression safety, mechanical cleanliness, test correctness, scope and code quality
20+ open-source developers handcrafted 150 tasks, each of which took over 40 hours to construct
it also tests a more diverse set of programming languages
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
SWE-Bench style grading has been the standard for years now - you ask the agent to solve an issue and then run its code on a pre-constructed unit test.
The problem is that passing a unit test is only one part of writing production-ready code. You also want to evaluate agents on a number of other axes, including scope, coding style, and unintended side effects.
The result is our new benchmark FrontierCode - which has ~80% fewer false positives and for which the best model (Opus 4.8) only scores 13%!
"Where others grade like a CI, FrontierCode grades like a tech lead."
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
congrats to Kimi on benchmaxxing this future benchmark to be on Sonnet 4.6 level (released 2 months prior to K2.6)
Mythos is live! so excited to have our FrontierCode recognized as the next frontier coding bench.
on FC Diamond, BOTH Opus 4.8 and GPT 5.5 don't meaningfully scale with effort, which many of you caught yesterday.
Mythos/Fable posttraining have really applied that test time compute toward solving very, very long running problems - dozens of human hour equivalents, hundreds of dollars per task, for the first time ever measured.
Available now in @Cognition @Devin for only 1.4x ACUs too! (I never thought i'd see this launch lol)
It's finally out!!! @METR_Evals found that more than half of SWEBench results is unmergeable slop. FrontierCode represents over 1000+ hours of maintainer validated software engineering work most frontier models cannot yet solve, much less solve with high quality.
Cog had IOI Gold medalists and top code maintainers Look At The Data — FrontierCode includes 3000+ rubrics covering code quality and anticheat reward hacking plaguing other benchmarks.
FC Diamond is so hard that Opus 4.8 scores 13.8%.
Three eras of AI coding : Three eras of benchmarks
2021 • Autocomplete : HumanEval 2023 • Passing Tests: SWEBench, TerminalBench 2026 • Maintainable Code: FrontierCode
to me the most beautiful chart when I requested a special historical run into all extant old models, the data was finding that the easiest third of FC tasks (in FC Extended) were rapidlly and suddenly solved over late 2025 - Opus almost doubled from a 41% pass rate to 74% in 4 months.
This describes the "WTF happened in Dec 2025" vibe shift that a lot of folks from @dhh to @karpathy have called out: it is the difference between getting 95% success in 2 rerolls vs 6, making it finally feasible to go up the next layer of abstraction in agentic coding, eg @GeoffreyHuntley's ralph loops or @bcherny's /goals or @steipete's "loops that prompt your agents" without fearing too much that things go off the rails.
My guess: as AI accelerates from here, each FrontierCode tier will saturate in sequence, hopefully ~annually. I've already asked the team to prepare FrontierCode 2027....
The old mountains will be destroyed. Their rubble becomes regolith. And from that regolith, the next model forest grows. Circle of life.
new frontier eval from the cognition team. interesting that simple test time scaling is pretty noisy here instead of a clean line
lots of care in crafting a good scoring process
https://cognition.ai/blog/frontier-code
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
20+ world-class open-source developers built realistic coding tasks on repos they maintain. They define what “mergeable” means in their repo.
What does it take to measure mergeability? We use a mix of unit tests, rubrics and novel verifiers to assess correctness, test quality, scope discipline, style, and adherence to codebase standards.
You can find full model results and technical implementation details on our blog:
https://cognition.ai/blog/frontier-code
A lot of people are already deploying AI into production codebases, but until now we didn’t really have a good eval for whether it writes code that is actually high-quality and maintainable. Pretty cool grading here:
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
Oh my God! @METR_Evals’s coding benchmarks are saturated! 🤯 Mythos broke the METR graph 🤯
4 weeks later, out comes a new coding task, this time from @cognition: “FrontierCode Diamond remains unsaturated: the best performing model, Claude Opus 4.8, achieves a score of only 13.4%.
There is still a lots of headroom.
*Note that METR itself never panicked. It’s the Twitterverse that has egg on its face.
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
FrontierCode has three task sets: Extended (150 tasks), Main (100 tasks) and Diamond (50 tasks). SOTA LLMs have significant room for improvement, with the top model earning a score of just 13.4/100 on our Diamond task set.
just finished rerunning FC Diamond on my historical charts. none of the official tables/charts are capturing the degree of takeoff.
its this same chart all the way down difficulty classes (below) breaks every curve fit because Fable is a diffferent CLASS of model, with beeeeeg model smell.
This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time.
I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!
FrontierCode was built in close partnership with the expert maintainers of 36 flagship open-source repositories, like @smilingnosrati, CEO & Tech Lead @CeleryOrg (29k stars), and Martin McKeaveney, CTO of @Budibase (28k stars).
Maintainers invested more than 40 hours per task, undergoing multiple rounds of iteration to ensure that any PR that satisfies these standards would actually be merged.
Rigorous quality control is important, so we built an extensive QC pipeline with adversarial testing, calibration, and multi-stage review. Every task is manually reviewed by a Cognition researcher.
This reduces both the false positive and false negative rates for proposed solutions. As a result, FrontierCode produces 81% fewer misclassification errors than SWE-Bench Pro.
Tasks in the dataset have a concise problem statement with large solutions that cut across multiple files.
FrontierCode’s task set is more diverse than other software engineering evals, measuring ability across a wide range of languages and problem types.
look at the data
It's finally out!!! @METR_Evals found that more than half of SWEBench results is unmergeable slop. FrontierCode represents over 1000+ hours of maintainer validated software engineering work most frontier models cannot yet solve, much less solve with high quality.
Cog had IOI Gold medalists and top code maintainers Look At The Data — FrontierCode includes 3000+ rubrics covering code quality and anticheat reward hacking plaguing other benchmarks.
FC Diamond is so hard that Opus 4.8 scores 13.8%.
Three eras of AI coding : Three eras of benchmarks
2021 • Autocomplete : HumanEval 2023 • Passing Tests: SWEBench, TerminalBench 2026 • Maintainable Code: FrontierCode
to me the most beautiful chart when I requested a special historical run into all extant old models, the data was finding that the easiest third of FC tasks (in FC Extended) were rapidlly and suddenly solved over late 2025 - Opus almost doubled from a 41% pass rate to 74% in 4 months.
This describes the "WTF happened in Dec 2025" vibe shift that a lot of folks from @dhh to @karpathy have called out: it is the difference between getting 95% success in 2 rerolls vs 6, making it finally feasible to go up the next layer of abstraction in agentic coding, eg @GeoffreyHuntley's ralph loops or @bcherny's /goals or @steipete's "loops that prompt your agents" without fearing too much that things go off the rails.
My guess: as AI accelerates from here, each FrontierCode tier will saturate in sequence, hopefully ~annually. I've already asked the team to prepare FrontierCode 2027....
The old mountains will be destroyed. Their rubble becomes regolith. And from that regolith, the next model forest grows. Circle of life.
The whole discourse about «the gap» is very annoying because we don't have an idea of the costs that go into making these gains. What is the compute delta between Kimi K2.5 and K2.6? Opus 4.7 and 4.8? Is the (in)efficiency of Cursor's Kimi post-training representative?
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
New hardest coding eval just dropped. Let’s see how long it takes to saturate.
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.
Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?
@cognition I have a few questions after looking deeper at the numbers