DeepSWE 1.0
DeepSWE 1.0 is Datacurve's long-horizon software-engineering benchmark for frontier coding agents, using 113 original tasks across 91 repositories and 5 languages. Model scores must preserve the evaluator and harness used for each run because xAI reports Grok 4.5's DeepSWE 1.0 score from a Datacurve-created eval run with each model provider's harnesses by AA, not the public Datacurve leaderboard table.
Models ranked
1
tracked on this benchmark
Score band
62.0 – 62.0
best → lowest tracked
Snapshot trend
—
need ≥2 snapshots
Leaderboard
Tracked models ranked by Pass@1 (higher is better).
Notes: xAI launch blog DeepSWE 1.0 chart reports Grok 4.5 at 62% Pass@1. Evaluator/source: Datacurve-created eval run by Artificial Analysis with each model provider's harness. Variant: Grok 4.5 base API model. Harness status: AA-run provider harness (xAI console/API path per blog caption); not the public Datacurve DeepSWE leaderboard table. Confidence: medium for vendor-reported chart value, medium-low for cross-model comparability. Recommended seed value: 62.
How to read this benchmark
This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.
Trust this score when
- There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
- The model list covers the same version family you can actually deploy today.
- Top candidates overlap with your required routing and feature requirements.
Be cautious when
- There is only one benchmark snapshot or the dataset appears stale.
- The benchmark metric direction is opposite of your decision objective.
- The score difference between options is narrow and likely within implementation variance.
Related benchmarks
Last reviewed: Jul 8, 2026