{"model": {"slug": "claude-4-5-opus-medium", "name": "Claude 4.5 Opus (medium)", "vendor": "Anthropic", "model_family": "claude-4-5-opus", "parameter_count_b": null, "context_length": 200000, "homepage_url": null, "hf_model_id": null, "openrouter_id": null, "card_summary": "Anthropic Claude 4.5 Opus (medium) on SWE-bench Verified leaderboards.", "is_rl_checkpoint": false, "score_count": 3, "best_success_rate": 79.2, "input_price_per_million": null, "output_price_per_million": null}, "scores": [{"success_rate": 79.2, "resolved_count": 396, "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://github.com/OpenAutoCoder/live-swe-agent", "observed_at": "2025-12-15T00:00:00+00:00", "harness": {"slug": "live-swe-agent", "name": "Live SWE Agent", "vendor": "OpenAutoCoder", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-5-opus-medium", "name": "Claude 4.5 Opus (medium)", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 74.4, "resolved_count": 372, "total_count": 500, "cost_usd": 0.721244997, "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 bash-only (mini-SWE-agent)", "source": "swe_bench_site", "source_url": "https://www.anthropic.com/news/claude-opus-4-5", "observed_at": "2025-11-24T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-5-opus-medium", "name": "Claude 4.5 Opus (medium)", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-bash-only", "name": "SWE-bench Verified (bash-only)"}}, {"success_rate": 74.4, "resolved_count": 372, "total_count": 500, "cost_usd": 0.721244997, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 38, "is_harnessrl_measured": false, "eval_protocol": "SWE-bench Verified public submission", "source": "swe_bench_site", "source_url": "https://mini-swe-agent.com/latest/", "observed_at": "2025-11-24T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-5-opus-medium", "name": "Claude 4.5 Opus (medium)", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}], "harnesses": [{"slug": "live-swe-agent", "name": "Live SWE Agent", "vendor": "OpenAutoCoder", "repo_url": "https://github.com/OpenAutoCoder/live-swe-agent", "openrouter_icon_url": null}, {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": "https://github.com/SWE-agent/mini-swe-agent", "openrouter_icon_url": null}]}