Global AI MapTop Harness Leaderboard
SWE-bench Verified: submitted agent systems solving real software issues. Scores are percent resolved; each entry retains its agent, model and checked status.
Build snapshot retrieved 2026-10-30T22:39:25.727891+00:00. Live data loads in your browser.
| RANK | SYSTEM / MODEL | SCORE | CONTEXT |
|---|---|---|---|
| 01 | Sonar Foundation Agent + Claude 4.5 Opus ↗ | 79.20 | SWE-bench Verified · Sonar Foundation Agent · submitted result 2025-12-05 |
| 02 | live-SWE-agent + Claude 4.5 Opus medium (20251101) ↗ | 79.20 | SWE-bench Verified · live-SWE-agent · submitted result 2025-12-15 |
| 03 | TRAE + Doubao-Seed-Code ↗ | 78.80 | SWE-bench Verified · TRAE · submitted result 2025-09-28 |
| 04 | live-SWE-agent + Gemini 3 Pro Preview (2025-11-18) ↗ | 77.40 | SWE-bench Verified · live-SWE-agent · checked run 2025-11-20 |
| 05 | EPAM AI/Run Developer Agent v20250719 + Claude 4 Sonnet ↗ | 76.80 | SWE-bench Verified · EPAM AI/Run Developer Agent · submitted result 2025-08-04 |
| 06 | Atlassian Rovo Dev (2025-09-02) ↗ | 76.80 | SWE-bench Verified · Atlassian Rovo Dev · submitted result 2025-09-02 |
| 07 | Claude 4.5 Opus (high) ↗ | 76.80 | SWE-bench Verified · mini-SWE-agent · submitted result 2026-02-17 |
| 08 | ACoder ↗ | 76.40 | SWE-bench Verified · ACoder · submitted result 2025-08-19 |
| 09 | Gemini 3 Flash (high) ↗ | 75.80 | SWE-bench Verified · mini-SWE-agent · submitted result 2026-02-17 |
| 10 | MiniMax M2.5 (high) ↗ | 75.80 | SWE-bench Verified · mini-SWE-agent · submitted result 2026-02-17 |
| 11 | Warp ↗ | 75.60 | SWE-bench Verified · Warp · submitted result 2025-09-01 |
| 12 | Claude 4.6 Opus ↗ | 75.60 | SWE-bench Verified · mini-SWE-agent · submitted result 2026-02-17 |
| 13 | TRAE + Claude Sonnet 4 + Opus 4 + Sonnet 3.7 + Gemini 2.5 Pro ↗ | 75.20 | SWE-bench Verified · TRAE · submitted result 2025-06-12 |
| 14 | Harness AI ↗ | 74.80 | SWE-bench Verified · Harness AI · submitted result 2025-07-31 |
| 15 | Sonar Foundation Agent + Claude 4.5 Sonnet ↗ | 74.80 | SWE-bench Verified · Sonar Foundation Agent · submitted result 2025-11-03 |
| 16 | Lingxi-v1.5_claude-4-sonnet-20250514 ↗ | 74.60 | SWE-bench Verified · Lingxi-v1.5 · submitted result 2025-07-20 |
| 17 | JoyCode + Claude 4 Sonnet + GPT-4.1 ↗ | 74.60 | SWE-bench Verified · JoyCode · submitted result 2025-09-15 |
| 18 | Refact.ai Agent + Claude 4 Sonnet + o4-mini ↗ | 74.40 | SWE-bench Verified · Refact.ai Agent · submitted result 2025-06-03 |
| 19 | Prometheus-v1.2.1 + GPT-5 ↗ | 74.40 | SWE-bench Verified · Prometheus-v1.2.1 · submitted result 2025-10-15 |
| 20 | Claude 4.5 Opus (20251101) (medium) ↗ | 74.40 | SWE-bench Verified · mini-SWE-agent · checked run 2025-11-24 |
Method, limits & original data
SWE-bench Verified: submitted agent systems solving real software issues. Scores are percent resolved; each entry retains its agent, model and checked status.
Rows are ranked within this benchmark. Different hardware, model versions, harnesses and test conditions can change results. A blank or absent score is never treated as zero. Ties retain equal rank in the interactive view.
Additional benchmark references: Arena · Artificial Analysis · Terminal-Bench. Keyed APIs are not exposed in this website.