Terminal-Bench 4.0.0 · Opus 5 / Claude Code

Created by Ryan Marten
||Owned by Terminal-Bench
PublicRerun job
330 / 330 finished
Started Aug 26, 2026, 6:18 PM · Ran for 24h 22m
324 completed6 errors0 in progress0 pending
harbor run -d terminal-bench/terminal-bench -a claude-code -m anthropic/claude-opus-5
{  "debug": false,  "quiet": false,  "retry": {    "max_retries": 3,    "max_wait_sec": 60,    "min_wait_sec": 1,    "wait_multiplier": 1,    "exclude_exceptions": [      "RewardFileNotFoundError",      "AgentTimeoutError",      "RewardFileEmptyError",      "AgentSafetyRefusalError",      "VerifierTimeoutError",      "ApiUsageLimitError",      "VerifierOutputParseError"    ]  },  "tasks": [],  "agents": [    {      "env": {        "CLAUDE_CODE_SIMPLE": "1"      },      "name": "claude-code",      "kwargs": {        "version": "2.1.231",        "reasoning_effort": "max"      },      "skills": [],      "model_name": "anthropic/claude-opus-5",      "mcp_servers": [],      "extra_allowed_hosts": []    }  ],  "metrics": [],  "datasets": [    {      "ref": "4",      "name": "terminal-bench/terminal-bench",      "overwrite": false    }  ],  "job_name": "Terminal-Bench 4.0.0 · Opus 5 / Claude Code",  "jobs_dir": "jobs",  "verifier": {    "disable": false  },  "artifacts": [],  "n_attempts": 5,  "environment": {    "type": "docker",    "delete": true,    "kwargs": {},    "force_build": false,    "extra_allowed_hosts": [],    "extra_docker_compose": [],    "cpu_enforcement_policy": "auto",    "memory_enforcement_policy": "auto"  },  "credential_mode": "direct",  "extra_instructions": [],  "timeout_multiplier": 1,  "n_concurrent_trials": 330,  "extra_instruction_paths": []}