{"harness": {"slug": "codesweep-swe-agent", "name": "CodeSweep SWE Agent KIMI K2 Instruct", "vendor": "CodeSweep Inc.", "description": "CodeSweep SWE Agent KIMI K2 Instruct is tracked in this catalog as a software-engineering agent harness. Public benchmark rows use this slug on leaderboards such as SWE-bench Verified or Artificial Analysis coding-agent evals.\n\nCheck each score row's `source` and `source_url` for the authoritative snapshot. Harness-level metadata here does not invent scores \u2014 only documents what the harness is and where published results come from.", "card_summary": "CodeSweep Inc. CodeSweep SWE Agent KIMI K2 Instruct coding agent harness.", "homepage_url": "https://codesweep.ai/", "repo_url": null, "docs_url": null, "scaffold_code": null, "harbor_agent_name": null, "api_protocol": null, "status": "validated", "supports_rl": true, "score_count": 1, "best_success_rate": 53.4, "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/20250804_codesweep_sweagent_kimi_k2_instruct.svg", "aa_coding_index": null, "aa_mean_cost_usd": null, "aa_mean_total_tokens": null}, "scores": [{"success_rate": 53.4, "resolved_count": 267, "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://codesweep.ai/", "observed_at": "2025-08-04T00:00:00+00:00", "harness": {"slug": "codesweep-swe-agent", "name": "CodeSweep SWE Agent KIMI K2 Instruct", "vendor": "CodeSweep Inc.", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "codesweep-sweagent-kimi-k2", "name": "Codesweep Sweagent KIMI K2", "vendor": "Moonshot"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}], "benchmarks": [{"slug": "swe-bench-verified", "name": "SWE-bench Verified", "category": "eval"}]}