{"model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic", "model_family": "claude-4-6-opus", "parameter_count_b": null, "context_length": 200000, "homepage_url": null, "hf_model_id": null, "openrouter_id": null, "card_summary": "Anthropic Claude 4.6 Opus on SWE-bench Verified leaderboards.", "is_rl_checkpoint": false, "score_count": 10, "best_success_rate": 80.0, "input_price_per_million": null, "output_price_per_million": null}, "scores": [{"success_rate": 80.0, "resolved_count": null, "total_count": null, "cost_usd": 0.44, "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": "OpenHands SDK v1.11.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/09286fea-ded4-4f32-a8b1-97e3a5847b83", "observed_at": "2026-07-31T11:38:58.577710", "harness": {"slug": "openhands-sdk", "name": "OpenHands SDK", "vendor": "All Hands AI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "gaia", "name": "GAIA"}}, {"success_rate": 78.8, "resolved_count": null, "total_count": null, "cost_usd": 0.43, "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": "OpenHands SDK v1.11.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/cbf7f857-21d0-4526-b830-25d4e88ab0ed", "observed_at": "2026-07-31T11:38:58.577710", "harness": {"slug": "openhands-sdk", "name": "OpenHands SDK", "vendor": "All Hands AI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swt-bench", "name": "SWT-Bench"}}, {"success_rate": 76.8, "resolved_count": null, "total_count": null, "cost_usd": 0.77, "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": "OpenHands SDK v1.11.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/77bb7715-dcff-4718-8e4f-611eae769107", "observed_at": "2026-07-31T11:38:58.577710", "harness": {"slug": "openhands-sdk", "name": "OpenHands SDK", "vendor": "All Hands AI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "openhands-swe-bench", "name": "SWE-bench (OpenHands Index)"}}, {"success_rate": 75.6, "resolved_count": 378, "total_count": 500, "cost_usd": 0.5515226445, "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/claude", "observed_at": "2026-02-17T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-bash-only", "name": "SWE-bench Verified (bash-only)"}}, {"success_rate": 75.6, "resolved_count": 378, "total_count": 500, "cost_usd": 0.5515226445, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 29, "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": "2026-02-17T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 72.0, "resolved_count": 216, "total_count": 300, "cost_usd": 0.6629735766666667, "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 Multilingual (mini-SWE-agent)", "source": "swe_bench_site", "source_url": "https://www.anthropic.com/news/claude-opus-4-5", "observed_at": "2026-02-13T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-multilingual", "name": "SWE-bench Multilingual"}}, {"success_rate": 63.11, "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": "MLE-bench All split any_medal_percentage", "source": "mle_bench_site", "source_url": "https://github.com/openai/mle-bench", "observed_at": "2026-03-06T00:00:00+00:00", "harness": {"slug": "aibuildai", "name": "AIBuildAI", "vendor": "Aibuildai", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "mle-bench", "name": "MLE-bench"}}, {"success_rate": 56.2, "resolved_count": null, "total_count": null, "cost_usd": 7.69, "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": "OpenHands SDK v1.11.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/8a792401-ab45-48b6-8398-99a1a7534f69", "observed_at": "2026-07-31T11:38:58.577710", "harness": {"slug": "openhands-sdk", "name": "OpenHands SDK", "vendor": "All Hands AI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "commit0", "name": "Commit0"}}, {"success_rate": 51.6, "resolved_count": null, "total_count": null, "cost_usd": 1.35, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 60, "is_harnessrl_measured": false, "eval_protocol": "WildClawBench overall score (60 tasks)", "source": "wildclawbench_site", "source_url": "https://github.com/internlm/WildClawBench", "observed_at": null, "harness": {"slug": "openclaw", "name": "OpenClaw", "vendor": "OpenClaw", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "wildclawbench", "name": "WildClawBench"}}, {"success_rate": 41.8, "resolved_count": null, "total_count": null, "cost_usd": 2.37, "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": "OpenHands SDK v1.11.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/0fe00101-1f0a-4698-8a43-70fd665f4936", "observed_at": "2026-07-31T11:38:58.577710", "harness": {"slug": "openhands-sdk", "name": "OpenHands SDK", "vendor": "All Hands AI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-6-opus", "name": "Claude 4.6 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-multimodal", "name": "SWE-bench Multimodal (OpenHands Index)"}}], "harnesses": [{"slug": "aibuildai", "name": "AIBuildAI", "vendor": "Aibuildai", "repo_url": "https://github.com/aibuildai/AI-Build-AI", "openrouter_icon_url": null}, {"slug": "openclaw", "name": "OpenClaw", "vendor": "OpenClaw", "repo_url": "https://github.com/openclaw/openclaw", "openrouter_icon_url": null}, {"slug": "openhands-sdk", "name": "OpenHands SDK", "vendor": "All Hands AI", "repo_url": "https://github.com/All-Hands-AI/OpenHands", "openrouter_icon_url": "https://www.all-hands.dev/favicon.ico"}, {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": "https://github.com/SWE-agent/mini-swe-agent", "openrouter_icon_url": null}]}