JobBench Leaderboard

Benchmark Model Rank Results
agents-on-jobbench-main-65-tasksClaude Opus 5.5 (OpenCode)#1Score: 75.4
agents-on-jobbench-main-65-tasksGPT-6 Astra (OpenCode)#2Score: 66.1
agents-on-jobbench-main-65-tasksClaude Opus 5 (OpenCode)#3Score: 65.7
agents-on-jobbench-main-65-tasksGLM-5.3 (OpenCode)#5Score: 61.4
agents-on-jobbench-main-65-tasksClaude Fable 5 (with fallbacks) (OpenCode)#6Score: 57.4
agents-on-jobbench-main-65-tasksMuse Spark 1.1 (OpenCode)#8Score: 54.7
agents-on-jobbench-main-65-tasksKimi K3 (OpenCode)#9Score: 54.3
agents-on-jobbench-main-65-tasksQwen3.8-Max (OpenCode)#10Score: 52.7
agents-on-jobbench-main-65-tasksGPT-5.6 Sol (OpenCode)#11Score: 45.4
agents-on-jobbench-main-65-tasksGLM-5.2 (OpenCode)#12Score: 43.4
agents-on-jobbench-main-65-tasksInkling (OpenCode)#13Score: 31.1
agents-on-jobbench-main-65-tasksNemotron 3 Ultra (OpenCode)#14Score: 27.2
agents-on-jobbench-main-65-tasksClaude Fable 5.1 (OpenCode)–Score: 73.1
agents-on-jobbench-main-65-tasksClaude Opus 4.7 (OpenCode)–Score: 44.5
agents-on-jobbench-main-65-tasksClaude Opus 4.8 (OpenCode)–Score: 48.4
agents-on-jobbench-main-65-tasksClaude Sonnet 4.5 (OpenCode)–Score: 20.7
agents-on-jobbench-main-65-tasksClaude Sonnet 4.6 (OpenCode)–Score: 36.6
agents-on-jobbench-main-65-tasksGPT-5.2 (OpenCode)–Score: 26.6
agents-on-jobbench-main-65-tasksGPT-5.4 (OpenCode)–Score: 32.2
agents-on-jobbench-main-65-tasksGPT-5.5 (OpenCode)–Score: 38.3
agents-on-jobbench-main-65-tasksGemini 3.1 Pro (OpenCode)–Score: 15.9
agents-on-jobbench-main-65-tasksGemini 3.5 Flash (OpenCode)–Score: 31.5
agents-on-jobbench-main-65-tasksMuse Spark 1.2 (OpenCode)–Score: 61.6
agents-on-jobbench-main-65-tasksMuse Spark 1.3 (OpenCode)–Score: 64.9