Software Engineering on SWE Verified HIGH reasoning
60.7Accuracygpt-oss-20b
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| gpt-oss-20bEvaluation Harness=OpenAI Published2026.04 | 60.7 | — | |
| HarmonyAgentEvaluation Harness=Harmony harness2026.04 | 60.4 | 56.2 |
| Method | Links | ||
|---|---|---|---|
| gpt-oss-20bEvaluation Harness=OpenAI Published2026.04 | 60.7 | — | |
| HarmonyAgentEvaluation Harness=Harmony harness2026.04 | 60.4 | 56.2 |