{"harness": {"slug": "tools", "name": "Tools", "vendor": "Anthropic", "description": "On the public SWE-bench Verified leaderboard, \"Tools\" is the harness name attached to several Anthropic Claude submissions that use a tool-using agent setup (distinct from the Claude Code product slug or bash-only mini-SWE-agent baselines). Rows pair this label with specific Claude snapshot dates in the model column.\n\nThis catalog ingests those community rows via `swe_bench_site`. The slug `tools` is a leaderboard artifact \u2014 not a separate open-source repository. For Anthropic's shipping CLI product, see `claude-code`; for minimal model comparisons, see `mini-swe-agent`.\n\nAll published scores here target `swe-bench-verified` (500 verified issues). Check each row's `source_url` for the submission snapshot on swebench.com.", "card_summary": "Anthropic \"Tools\" harness label on SWE-bench Verified \u2014 Claude models with tool-use agent configuration.", "homepage_url": "https://www.anthropic.com", "repo_url": "https://github.com/anthropics", "docs_url": null, "scaffold_code": null, "harbor_agent_name": null, "api_protocol": null, "status": "validated", "supports_rl": true, "score_count": 5, "best_success_rate": 73.2, "github_stars": null, "popularity_tier": null, "hrl_score": null, "hrl_rank": null, "catalog_token_volume": null, "openrouter_icon_url": null, "leaderboard_icon_url": "https://www.swebench.com/img/logos/20241022_tools_claude-3-5-haiku.jpg", "aa_coding_index": null, "aa_mean_cost_usd": null, "aa_mean_total_tokens": null}, "scores": [{"success_rate": 73.2, "resolved_count": 366, "total_count": 500, "cost_usd": null, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": null, "is_harnessrl_measured": false, "eval_protocol": "SWE-bench Verified public submission", "source": "swe_bench_site", "source_url": "https://www.anthropic.com/", "observed_at": "2025-05-22T00:00:00+00:00", "harness": {"slug": "tools", "name": "Tools", "vendor": "Anthropic", "repo_url": "https://github.com/anthropics", "openrouter_icon_url": null}, "model": {"slug": "claude-4-opus", "name": "Claude 4 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 72.4, "resolved_count": 362, "total_count": 500, "cost_usd": null, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": null, "is_harnessrl_measured": false, "eval_protocol": "SWE-bench Verified public submission", "source": "swe_bench_site", "source_url": "https://www.anthropic.com/", "observed_at": "2025-05-22T00:00:00+00:00", "harness": {"slug": "tools", "name": "Tools", "vendor": "Anthropic", "repo_url": "https://github.com/anthropics", "openrouter_icon_url": null}, "model": {"slug": "claude-4-sonnet", "name": "Claude 4 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 63.2, "resolved_count": 316, "total_count": 500, "cost_usd": null, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": null, "is_harnessrl_measured": false, "eval_protocol": "SWE-bench Verified public submission", "source": "swe_bench_site", "source_url": "https://www.anthropic.com/", "observed_at": "2025-02-24T00:00:00+00:00", "harness": {"slug": "tools", "name": "Tools", "vendor": "Anthropic", "repo_url": "https://github.com/anthropics", "openrouter_icon_url": null}, "model": {"slug": "claude-3-7-sonnet-20250224", "name": "Claude 3.7 Sonnet (2025-02-24)", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 49.0, "resolved_count": 245, "total_count": 500, "cost_usd": null, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": null, "is_harnessrl_measured": false, "eval_protocol": "SWE-bench Verified public submission", "source": "swe_bench_site", "source_url": "https://www.anthropic.com/", "observed_at": "2024-10-22T00:00:00+00:00", "harness": {"slug": "tools", "name": "Tools", "vendor": "Anthropic", "repo_url": "https://github.com/anthropics", "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet-20241022", "name": "Claude 3.5 Sonnet (2024-10-22)", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 40.6, "resolved_count": 203, "total_count": 500, "cost_usd": null, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": null, "is_harnessrl_measured": false, "eval_protocol": "SWE-bench Verified public submission", "source": "swe_bench_site", "source_url": "https://www.anthropic.com/", "observed_at": "2024-10-22T00:00:00+00:00", "harness": {"slug": "tools", "name": "Tools", "vendor": "Anthropic", "repo_url": "https://github.com/anthropics", "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-haiku", "name": "Claude 3.5 Haiku", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}], "benchmarks": [{"slug": "swe-bench-verified", "name": "SWE-bench Verified", "category": "eval"}]}