Software Engineering on SWE Verified MEDIUM reasoning
53.3Overall ScoreHarmonyAgent
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HarmonyAgentEvaluation Harness=Harmony harness2026.04 | 53.3 | 49.3 | |
| gpt-oss-20bEvaluation Harness=OpenAI Published2026.04 | 53.2 | — |
| Method | Links | ||
|---|---|---|---|
| HarmonyAgentEvaluation Harness=Harmony harness2026.04 | 53.3 | 49.3 | |
| gpt-oss-20bEvaluation Harness=OpenAI Published2026.04 | 53.2 | — |