{"harness": {"slug": "nebius-ai", "name": "Nebius AI QWEN 2.5 72B Generator", "vendor": "Nebius", "description": "Nebius AI QWEN 2.5 72B Generator 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": "Nebius Nebius AI QWEN 2.5 72B Generator coding agent harness.", "homepage_url": "https://nebius.com/blog/posts/training-and-search-for-software-engineering-agents", "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": 40.6, "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/20241113_nebius-search-open-weight-models-11-24.png", "aa_coding_index": null, "aa_mean_cost_usd": null, "aa_mean_total_tokens": null}, "scores": [{"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://nebius.com/blog/posts/training-and-search-for-software-engineering-agents", "observed_at": "2024-11-13T00:00:00+00:00", "harness": {"slug": "nebius-ai", "name": "Nebius AI QWEN 2.5 72B Generator", "vendor": "Nebius", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "ensemble-nebius-ai-qwen-2-5-72b-generator-llama-3-1-70b-critic", "name": "Ensemble (Nebius AI QWEN 2.5 72B Generator, Llama 3.1 70B Critic)", "vendor": "Qwen"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}], "benchmarks": [{"slug": "swe-bench-verified", "name": "SWE-bench Verified", "category": "eval"}]}