Benchmarks and leaderboards

SWE-Bench Pro ranking as of 2026-10-07: Scale AI's official public and private dataset boards, verbatim

As of 2026-10-07, Scale AI's official SWE-Bench Pro public board lists Muse Spark 1.1* at 61.50 and gpt-5.4 (xHigh)* at 59.10. Board text only, no model picks.

Updated 2026-10-08

Evaluating Ornith-1.5 yourself

Judge a new model against your own work instead of public leaderboards.

Updated 2026-10-08

AA Index Explained: v4.3.2 Weights and 10 Evals

The 10 evaluations and category weights as of 2026-10-07, compared with the old v4.1 weights

Updated 2026-10-07

Terminal-Bench 3.0 Explained

TB 3.0 vs. 2.1, current rankings, and which leaders are sellable

Updated 2026-09-21

Terminal-Bench 2.1 Leaderboard Explained

Three boards, three winners: the official 83.8 / AA 89.5 / vals 85.77 gap fully explained

Updated 2026-09-21

SWE-bench Pro's Three First Places Explained

80.3 / 80.0 / the official board: the truth about data splits and scaffolds

Updated 2026-08-22

From SWE-Bench to Production: How Benchmarks Do (Not) Translate to Delivery Speed in 2026

88% on SWE-Bench sounds great, but often only 30% on your private repos. Understand the gap and build your own verification harness.

Updated 2026-08-17

Claude Fable 5 SWE-Bench Pro Score, Explained

Where the 80.3% figure comes from, official methodology vs. independent reproduction, and how much it should weigh in your decision

Updated 2026-08-15