{"model": {"slug": "claude-3-opus", "name": "Claude 3 Opus", "vendor": "Anthropic", "model_family": "claude-3-opus", "parameter_count_b": null, "context_length": 200000, "homepage_url": null, "hf_model_id": null, "openrouter_id": null, "card_summary": "Anthropic Claude 3 Opus on SWE-bench Verified leaderboards.", "is_rl_checkpoint": false, "score_count": 3, "best_success_rate": 15.8, "input_price_per_million": null, "output_price_per_million": null}, "scores": [{"success_rate": 15.8, "resolved_count": 79, "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/SWE-agent/SWE-agent", "observed_at": "2024-04-02T00:00:00+00:00", "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-opus", "name": "Claude 3 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 11.67, "resolved_count": 35, "total_count": 300, "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 Lite public submission", "source": "swe_bench_site", "source_url": "https://github.com/SWE-agent/SWE-agent", "observed_at": "2024-04-02T00:00:00+00:00", "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-opus", "name": "Claude 3 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-lite", "name": "SWE-bench Lite"}}, {"success_rate": 10.51, "resolved_count": null, "total_count": null, "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 Test public submission", "source": "swe_bench_site", "source_url": "https://www.swebench.com/", "observed_at": "2024-04-02T00:00:00+00:00", "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-opus", "name": "Claude 3 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-test", "name": "SWE-bench Test"}}], "harnesses": [{"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": "https://github.com/SWE-agent/SWE-agent", "openrouter_icon_url": null}]}