{"model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic", "model_family": "claude-3-5-sonnet", "parameter_count_b": null, "context_length": 200000, "homepage_url": null, "hf_model_id": null, "openrouter_id": null, "card_summary": "Anthropic Claude 3.5 Sonnet (leaderboard naming variant).", "is_rl_checkpoint": false, "score_count": 12, "best_success_rate": 60.2, "input_price_per_million": null, "output_price_per_million": null}, "scores": [{"success_rate": 60.2, "resolved_count": 301, "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://arxiv.org/pdf/2501.10893", "observed_at": "2025-01-10T00:00:00+00:00", "harness": {"slug": "learn-by-interact", "name": "Learn By Interact", "vendor": "Google", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 47.0, "resolved_count": 141, "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/darsagent/DARS-Agent", "observed_at": "2025-02-05T00:00:00+00:00", "harness": {"slug": "dars-agent", "name": "DARS Agent", "vendor": "Darsagent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-lite", "name": "SWE-bench Lite"}}, {"success_rate": 44.21, "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-rebench resolved rate (best rolling window)", "source": "swe_rebench_site", "source_url": "https://swe-rebench.com/", "observed_at": null, "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-rebench", "name": "SWE-rebench"}}, {"success_rate": 39.6, "resolved_count": 198, "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.epam.com/services/artificial-intelligence", "observed_at": "2024-10-29T00:00:00+00:00", "harness": {"slug": "epam-ai-run-developer-agent", "name": "EPAM AI Run", "vendor": "EPAM Systems, Inc.", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 33.6, "resolved_count": 168, "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": "2024-06-20T00:00:00+00:00", "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 26.67, "resolved_count": 80, "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/aorwall/moatless-tools", "observed_at": "2024-06-23T00:00:00+00:00", "harness": {"slug": "moatless-tools", "name": "Moatless Tools", "vendor": "Aorwall", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-lite", "name": "SWE-bench Lite"}}, {"success_rate": 23.0, "resolved_count": 69, "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": "2024-06-20T00:00:00+00:00", "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-lite", "name": "SWE-bench Lite"}}, {"success_rate": 18.13, "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": "2024-06-20T00:00:00+00:00", "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-test", "name": "SWE-bench Test"}}, {"success_rate": 12.19, "resolved_count": 63, "total_count": 517, "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 Multimodal public submission", "source": "swe_bench_site", "source_url": "https://github.com/SWE-agent/SWE-agent", "observed_at": "2024-10-06T00:00:00+00:00", "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-multimodal-site", "name": "SWE-bench Multimodal (site)"}}, {"success_rate": 11.99, "resolved_count": 62, "total_count": 517, "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 Multimodal public submission", "source": "swe_bench_site", "source_url": "https://github.com/SWE-agent/SWE-agent", "observed_at": "2024-10-06T00:00:00+00:00", "harness": {"slug": "swe-agent-javascript", "name": "SWE-agent (JavaScript)", "vendor": "Princeton NLP", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-multimodal-site", "name": "SWE-bench Multimodal (site)"}}, {"success_rate": 11.41, "resolved_count": 59, "total_count": 517, "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 Multimodal public submission", "source": "swe_bench_site", "source_url": "https://github.com/SWE-agent/SWE-agent", "observed_at": "2024-10-06T00:00:00+00:00", "harness": {"slug": "swe-agent-multimodal", "name": "SWE-agent (multimodal)", "vendor": "Princeton NLP", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-multimodal-site", "name": "SWE-bench Multimodal (site)"}}, {"success_rate": 6.19, "resolved_count": 32, "total_count": 517, "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 Multimodal public submission", "source": "swe_bench_site", "source_url": "https://github.com/OpenAutoCoder/Agentless", "observed_at": "2024-10-06T00:00:00+00:00", "harness": {"slug": "agentless-lite", "name": "Agentless Lite", "vendor": "OpenAutoCoder", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-multimodal-site", "name": "SWE-bench Multimodal (site)"}}], "harnesses": [{"slug": "agentless-lite", "name": "Agentless Lite", "vendor": "OpenAutoCoder", "repo_url": "https://github.com/OpenAutoCoder/Agentless", "openrouter_icon_url": null}, {"slug": "dars-agent", "name": "DARS Agent", "vendor": "Darsagent", "repo_url": "https://github.com/darsagent/DARS-Agent", "openrouter_icon_url": null}, {"slug": "epam-ai-run-developer-agent", "name": "EPAM AI Run", "vendor": "EPAM Systems, Inc.", "repo_url": "https://github.com/epam/AI-RUN", "openrouter_icon_url": null}, {"slug": "learn-by-interact", "name": "Learn By Interact", "vendor": "Google", "repo_url": null, "openrouter_icon_url": null}, {"slug": "moatless-tools", "name": "Moatless Tools", "vendor": "Aorwall", "repo_url": "https://github.com/aorwall/moatless-tools", "openrouter_icon_url": null}, {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": "https://github.com/SWE-agent/SWE-agent", "openrouter_icon_url": null}, {"slug": "swe-agent-javascript", "name": "SWE-agent (JavaScript)", "vendor": "Princeton NLP", "repo_url": "https://github.com/SWE-agent/SWE-agent", "openrouter_icon_url": null}, {"slug": "swe-agent-multimodal", "name": "SWE-agent (multimodal)", "vendor": "Princeton NLP", "repo_url": "https://github.com/SWE-agent/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}]}