Charley Lee

co-founder @datacurve

[Charley Lee @charleyslee]

A very impressive open model

[Datacurve @datacurve]

GLM 5.2 is now on DeepSWE as the top open-source model on our leaderboard.

With a pass@1 score of 44% at max effort, GLM 5.2 is indisputable #1 open-source model besting Kimi K2.7 Code by 17%.

[Datacurve @datacurve]

Claude Fable 5 debuts at #1 on DeepSWE.

It outscores the previous best by 3% and sets a new state-of-the-art on our long-horizon coding benchmark.

[Artificial Analysis @ArtificialAnlys]

We've updated the Artificial Analysis Coding Agent Index, replacing SWE-Bench Pro with Datacurve's DeepSWE benchmark - the swap lifts Codex with GPT-5.5 (xhigh) above Claude Code with Opus 4.8 (max), while the newly released Claude Fable 5 (max) in Claude Code debuts at the top

DeepSWE, built by @datacurve, writes its tasks from scratch rather than adapting them from public GitHub issues or pull requests, so no model has seen the solutions during training. That matters because SWE-Bench Pro, the benchmark it replaces in our Coding Agent Index, had grown gameable, with some models recovering the fix from the repository's commit history instead of solving the task.

The swap reorders the index: Codex with GPT-5.5 (xhigh) rises from 65 to 76, overtaking Claude Code with Opus 4.8 (max) at 73. Claude Code with Fable 5 (max), which enters directly on the refreshed index, leads at 77. SWE-Bench Pro had been flattering some combinations and penalizing others.

More below.

[Charley Lee @charleyslee]

@elithrar Looks super interesting for our use case, dropped you a dm :)

[Charley Lee @charleyslee]

@chribjel @MatthewBerman @datacurve 👀

[Theo - t3.gg @theo]

swe-bench is kind of a shitshow, and it makes evaluating LLMs hard. DeepSWE is the first agentic code bench that makes sense.

[Charley Lee @charleyslee]

Our results are out for 4.8 on DeepSWE – overall better performance while being more efficient than 4.7

[Datacurve @datacurve]

Opus 4.8 is now on DeepSWE.

On the default high thinking effort, it scores 6% higher than Opus 4.7 xhigh, while also lowering average cost per task.

[Theo - t3.gg @theo]

This is the first code bench that actually aligns with how it feels to use these models coding.

[Charley Lee @charleyslee]

Excited to launch this as a measure of performance on truly uncontaminated SWE tasks

[Serena Ge (Datacurve) @serenaa_ge]

Today we’re releasing DeepSWE, a new standard for agentic coding benchmarks.

On public leaderboards, top models often look relatively close in capability. DeepSWE shows where they actually diverge, reflecting the realistic experience of developers in their day-to-day work.

[Charley Lee @charleyslee]

So many coding agent interfaces treat great code review UX as an afterthought (understandable for CLIs, but even GUIs??) when it’s such a critical part of the process

[Charley Lee @charleyslee]

@bentlegen Being able to comment inline like revdiff would be amazing 🥹

[Charley Lee @charleyslee]

@nicole_clash right but that's still an intelligence bottleneck..

[TBPN @tbpn]

NEWS: @datacurve raises $15M Series A, led by Chemistry

[Charley Lee @charleyslee]

@MelkeyDev 🤩

[Charley Lee @charleyslee]

It's been a bit since we started as UncleGPT

[Y Combinator @ycombinator]

Congrats to @datacurve on their $15M Series A!

Datacurve provides frontier training data to the world’s leading foundation-model labs, helping to push the boundaries of what AI can do.

https://t.co/8SCyvZ6OEt

[Serena Ge (Datacurve) @serenaa_ge]

Today we’re announcing we’ve raised $17.5 million in funding across a $15M Series A led by Chemistry and a $2.7M Seed to accelerate foundation model progress through providing frontier training data for LLMs.

When we first started Datacurve, it came from a simple realization: foundation model progress is limited not just by compute, but by data quality and complexity.

The right data unlocks new capabilities, especially in coding, where accuracy and reasoning matter most.

We’re now proud to partner with the world’s leading foundation-model labs, providing them with high-quality, complex training data that helps push the boundaries of what AI can do.

This is still just the start. Come build the future of technology with us in San Francisco: https://t.co/6vVsjUmj4t

Huge thanks to our incredible team and investors who’ve believed in us since day one and beyond: @garrytan at @ycombinator, @1vnzh from @cohere , @Mark_Goldberg_ from @chemistry_fund, @TheDerrickLi from @AforeVC, @forwarddeploy, @SoheilK, and @shyamalanadkat.

[Charley Lee @charleyslee]

@aidenybai Just started taking it. Can feel a little effectiveness I’d say

[Charley Lee @charleyslee]

@0xluffy Over karaoke, smh

[Charley Lee @charleyslee]

@0xluffy That’s messed up