DeepSWE 1.1 S
AI benchmark · Coding
What is DeepSWE 1.1?
DeepSWE 1.1 software-engineering benchmark comparison reported in xAI's Grok 4.5 launch post, using the mini-SWE-agent harness run by Datacurve.
Tier S
The standard: frontier labs report it, or a trusted evaluator scores hundreds of models on it.
How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.
- 58 models scored
- reported by xAI
- in curated RL directories (1 pt)
On X · mixed, moderate buzz, from 11 posts about DeepSWE 1.1 in the last month
Widely cited in lab launch comparisons and model results, but users report high run-to-run variance, harness sensitivity, and discovered reward hacks.
How we got to tier S
- Adoption: 58 models scored; reported by xAI. Tier S.
- On X: mixed, moderate buzz. “Widely cited in lab launch comparisons and model results, but users report high run-to-run variance, harness sensitivity, and discovered reward hacks.” No move.
- In curated RL directories: 1 point, a floor of none.
- Not yet confirmed by a line from its own site or paper.
A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.
Not on the map yet: no line from its own site or paper places it in a domain.
Next: browse all benchmarks. Entry from the RL Research daily scrape of public sources, 2026-10-11 snapshot.