{"benchmark": {"slug": "terminal-bench-2-1-official", "name": "Terminal-Bench 2.1 (official)", "description": "Official Harbor Terminal-Bench 2.1 agent submissions.", "category": "eval", "split": null, "language": "bash", "task_count": null, "homepage_url": "https://github.com/harbor-framework/terminal-bench-2-1", "paper_url": null, "hf_dataset_id": null, "official_leaderboard_url": "https://github.com/harbor-framework/terminal-bench-2-1", "is_public_leaderboard": true, "harness_count": 6, "model_count": 11, "score_count": 16, "best_success_rate": 83.82}, "protocol": null, "rankings": {"benchmark": {"slug": "terminal-bench-2-1-official", "name": "Terminal-Bench 2.1 (official)", "description": "Official Harbor Terminal-Bench 2.1 agent submissions.", "category": "eval", "split": null, "language": "bash", "task_count": null, "homepage_url": "https://github.com/harbor-framework/terminal-bench-2-1", "paper_url": null, "hf_dataset_id": null, "official_leaderboard_url": "https://github.com/harbor-framework/terminal-bench-2-1", "is_public_leaderboard": true, "harness_count": null, "model_count": null, "score_count": null, "best_success_rate": null}, "sort": "success_rate", "rows": [{"rank": 1, "harness": {"slug": "claude-code", "name": "Claude Code", "vendor": "Anthropic", "scaffold_code": "cc", "repo_url": "https://github.com/anthropics/claude-code", "homepage_url": "https://github.com/anthropics/claude-code", "openrouter_icon_url": "https://claude.ai/favicon.ico", "leaderboard_icon_url": null}, "model": {"slug": "fable-5", "name": "Fable 5", "vendor": "Anthropic"}, "success_rate": 83.82, "resolved_count": null, "total_count": null, "cost_usd": 1.2419550561797752, "input_tokens": 20480925, "output_tokens": 9954945, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/75", "observed_at": "2026-06-07T00:00:00+00:00"}, {"rank": 2, "harness": {"slug": "codex", "name": "Codex", "vendor": "OpenAI", "scaffold_code": null, "repo_url": "https://github.com/openai/codex", "homepage_url": "https://openai.com/codex", "openrouter_icon_url": "https://openai.com/favicon.ico", "leaderboard_icon_url": null}, "model": {"slug": "gpt-5-5", "name": "GPT-5.5", "vendor": "openai"}, "success_rate": 83.15, "resolved_count": null, "total_count": null, "cost_usd": 4.627393258426967, "input_tokens": 336797311, "output_tokens": 5966373, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/45", "observed_at": "2026-05-01T00:00:00+00:00"}, {"rank": 3, "harness": {"slug": "terminus-2", "name": "Terminus 2", "vendor": "Harbor", "scaffold_code": "terminus", "repo_url": "https://github.com/Elvin-Yiming-Du/harbor", "homepage_url": "https://github.com/Elvin-Yiming-Du/harbor", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "fable-5", "name": "Fable 5", "vendor": "Anthropic"}, "success_rate": 80.45, "resolved_count": null, "total_count": null, "cost_usd": 0.9857078651685393, "input_tokens": 7874199, "output_tokens": 7412519, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/78", "observed_at": "2026-06-05T00:00:00+00:00"}, {"rank": 4, "harness": {"slug": "cursor-cli", "name": "Cursor CLI", "vendor": "Cursor", "scaffold_code": null, "repo_url": "https://github.com/getcursor/cursor", "homepage_url": "https://cursor.com", "openrouter_icon_url": "https://cursor.com/apple-touch-icon.png", "leaderboard_icon_url": null}, "model": {"slug": "grok-4-5", "name": "GROK 4.5", "vendor": "x-ai"}, "success_rate": 79.33, "resolved_count": null, "total_count": null, "cost_usd": 0.3013258426966292, "input_tokens": 12570445, "output_tokens": 4734062, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/86", "observed_at": "2026-07-09T00:00:00+00:00"}, {"rank": 5, "harness": {"slug": "claude-code", "name": "Claude Code", "vendor": "Anthropic", "scaffold_code": "cc", "repo_url": "https://github.com/anthropics/claude-code", "homepage_url": "https://github.com/anthropics/claude-code", "openrouter_icon_url": "https://claude.ai/favicon.ico", "leaderboard_icon_url": null}, "model": {"slug": "claude-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "success_rate": 78.88, "resolved_count": null, "total_count": null, "cost_usd": 0.6448089887640449, "input_tokens": 12873472, "output_tokens": 8089069, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/92", "observed_at": "2026-07-09T00:00:00+00:00"}, {"rank": 6, "harness": {"slug": "codex", "name": "Codex", "vendor": "OpenAI", "scaffold_code": null, "repo_url": "https://github.com/openai/codex", "homepage_url": "https://openai.com/codex", "openrouter_icon_url": "https://openai.com/favicon.ico", "leaderboard_icon_url": null}, "model": {"slug": "gpt-5-6", "name": "GPT-5.6", "vendor": "OpenAI"}, "success_rate": 78.43, "resolved_count": null, "total_count": null, "cost_usd": 0.9464044943820225, "input_tokens": 29803660, "output_tokens": 8707506, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/115", "observed_at": "2026-07-11T00:00:00+00:00"}, {"rank": 7, "harness": {"slug": "terminus-2", "name": "Terminus 2", "vendor": "Harbor", "scaffold_code": "terminus", "repo_url": "https://github.com/Elvin-Yiming-Du/harbor", "homepage_url": "https://github.com/Elvin-Yiming-Du/harbor", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gpt-5-5", "name": "GPT-5.5", "vendor": "openai"}, "success_rate": 77.98, "resolved_count": null, "total_count": null, "cost_usd": 1.1097752808988766, "input_tokens": 81469069, "output_tokens": 10768129, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/47", "observed_at": "2026-05-01T00:00:00+00:00"}, {"rank": 8, "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "scaffold_code": null, "repo_url": "https://github.com/SWE-agent/mini-swe-agent", "homepage_url": "https://github.com/SWE-agent/mini-swe-agent", "openrouter_icon_url": null, "leaderboard_icon_url": "https://www.swebench.com/img/mini-icon.svg"}, "model": {"slug": "muse-spark-1-1", "name": "Muse Spark 1.1", "vendor": "meta"}, "success_rate": 76.18, "resolved_count": null, "total_count": null, "cost_usd": 0.44505617977528095, "input_tokens": 60996, "output_tokens": 13679953, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/94", "observed_at": "2026-07-09T00:00:00+00:00"}, {"rank": 9, "harness": {"slug": "claude-code", "name": "Claude Code", "vendor": "Anthropic", "scaffold_code": "cc", "repo_url": "https://github.com/anthropics/claude-code", "homepage_url": "https://github.com/anthropics/claude-code", "openrouter_icon_url": "https://claude.ai/favicon.ico", "leaderboard_icon_url": null}, "model": {"slug": "claude-5-sonnet", "name": "Claude 5 Sonnet", "vendor": "Anthropic"}, "success_rate": 74.61, "resolved_count": null, "total_count": null, "cost_usd": 0.6475955056179775, "input_tokens": 20137729, "output_tokens": 11216322, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/98", "observed_at": "2026-07-09T00:00:00+00:00"}, {"rank": 10, "harness": {"slug": "terminus-2", "name": "Terminus 2", "vendor": "Harbor", "scaffold_code": "terminus", "repo_url": "https://github.com/Elvin-Yiming-Du/harbor", "homepage_url": "https://github.com/Elvin-Yiming-Du/harbor", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gemini-3-pro", "name": "Gemini 3 Pro", "vendor": "Google"}, "success_rate": 73.93, "resolved_count": null, "total_count": null, "cost_usd": 0.5043595505617977, "input_tokens": 26094316, "output_tokens": 12862730, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/48", "observed_at": "2026-05-01T00:00:00+00:00"}, {"rank": 11, "harness": {"slug": "claude-code", "name": "Claude Code", "vendor": "Anthropic", "scaffold_code": "cc", "repo_url": "https://github.com/anthropics/claude-code", "homepage_url": "https://github.com/anthropics/claude-code", "openrouter_icon_url": "https://claude.ai/favicon.ico", "leaderboard_icon_url": null}, "model": {"slug": "claude-4-7-opus", "name": "Claude 4.7 Opus", "vendor": "Anthropic"}, "success_rate": 68.9, "resolved_count": null, "total_count": null, "cost_usd": 1.3412080536912752, "input_tokens": 31873388, "output_tokens": 14896548, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/44", "observed_at": "2026-05-01T00:00:00+00:00"}, {"rank": 12, "harness": {"slug": "terminus-2", "name": "Terminus 2", "vendor": "Harbor", "scaffold_code": "terminus", "repo_url": "https://github.com/Elvin-Yiming-Du/harbor", "homepage_url": "https://github.com/Elvin-Yiming-Du/harbor", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-4-7-opus", "name": "Claude 4.7 Opus", "vendor": "Anthropic"}, "success_rate": 66.07, "resolved_count": null, "total_count": null, "cost_usd": 1.3084494382022471, "input_tokens": 16646631, "output_tokens": 12451178, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/46", "observed_at": "2026-05-01T00:00:00+00:00"}, {"rank": 13, "harness": {"slug": "gemini-cli", "name": "Gemini CLI", "vendor": "Google", "scaffold_code": null, "repo_url": "https://github.com/google-gemini/gemini-cli", "homepage_url": "https://github.com/google-gemini/gemini-cli", "openrouter_icon_url": "https://www.gstatic.com/lamda/images/gemini_favicon_f06995846fc77c897+24.png", "leaderboard_icon_url": null}, "model": {"slug": "gemini-3-pro", "name": "Gemini 3 Pro", "vendor": "Google"}, "success_rate": 65.84, "resolved_count": null, "total_count": null, "cost_usd": 0.5567640449438203, "input_tokens": 47653156, "output_tokens": 7219388, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/66", "observed_at": "2026-05-01T00:00:00+00:00"}, {"rank": 14, "harness": {"slug": "gemini-cli", "name": "Gemini CLI", "vendor": "Google", "scaffold_code": null, "repo_url": "https://github.com/google-gemini/gemini-cli", "homepage_url": "https://github.com/google-gemini/gemini-cli", "openrouter_icon_url": "https://www.gstatic.com/lamda/images/gemini_favicon_f06995846fc77c897+24.png", "leaderboard_icon_url": null}, "model": {"slug": "gemini-3-1-pro", "name": "Gemini 3.1 Pro", "vendor": "Google"}, "success_rate": 65.84, "resolved_count": null, "total_count": null, "cost_usd": 0.5314382022471911, "input_tokens": 44843158, "output_tokens": 6123596, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/68", "observed_at": "2026-05-05T00:00:00+00:00"}, {"rank": 15, "harness": {"slug": "terminus-2", "name": "Terminus 2", "vendor": "Harbor", "scaffold_code": "terminus", "repo_url": "https://github.com/Elvin-Yiming-Du/harbor", "homepage_url": "https://github.com/Elvin-Yiming-Du/harbor", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gemini-3-1-pro", "name": "Gemini 3.1 Pro", "vendor": "Google"}, "success_rate": 65.62, "resolved_count": null, "total_count": null, "cost_usd": 0.5168314606741573, "input_tokens": 27525560, "output_tokens": 12935831, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/69", "observed_at": "2026-05-05T00:00:00+00:00"}, {"rank": 16, "harness": {"slug": "claude-code", "name": "Claude Code", "vendor": "Anthropic", "scaffold_code": "cc", "repo_url": "https://github.com/anthropics/claude-code", "homepage_url": "https://github.com/anthropics/claude-code", "openrouter_icon_url": "https://claude.ai/favicon.ico", "leaderboard_icon_url": null}, "model": {"slug": "glm-5-1", "name": "GLM 5 1", "vendor": "Zhipu"}, "success_rate": 58.65, "resolved_count": null, "total_count": null, "cost_usd": 0.6227865168539326, "input_tokens": 13570352, "output_tokens": 10635681, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/67", "observed_at": "2026-05-01T00:00:00+00:00"}]}}