{"model": {"slug": "claude-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic", "model_family": "claude-4-8-opus", "parameter_count_b": null, "context_length": 200000, "homepage_url": null, "hf_model_id": null, "openrouter_id": null, "card_summary": "Anthropic Claude 4.8 Opus on SWE-bench Verified leaderboards.", "is_rl_checkpoint": false, "score_count": 7, "best_success_rate": 84.3, "input_price_per_million": null, "output_price_per_million": null}, "scores": [{"success_rate": 84.3, "resolved_count": null, "total_count": null, "cost_usd": 0.73, "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 v1.18.1", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/efde50d3-3151-48fc-aac2-c98a80d7406e", "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-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swt-bench", "name": "SWT-Bench"}}, {"success_rate": 83.8, "resolved_count": null, "total_count": null, "cost_usd": 0.75, "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 v1.18.1", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/39ca1805-3f89-4219-8cdd-b6ca2360fb17", "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-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "openhands-swe-bench", "name": "SWE-bench (OpenHands Index)"}}, {"success_rate": 78.88, "resolved_count": null, "total_count": null, "cost_usd": 0.6448089887640449, "input_tokens": 12873472, "output_tokens": 8089069, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 445, "is_harnessrl_measured": false, "eval_protocol": null, "source": "tbench_official", "source_url": "https://github.com/harbor-framework/terminal-bench-2-1/pull/92", "observed_at": "2026-07-09T00:00:00+00:00", "harness": {"slug": "claude-code", "name": "Claude Code", "vendor": "Anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "claude-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "terminal-bench-2-1-official", "name": "Terminal-Bench 2.1 (official)"}}, {"success_rate": 78.8, "resolved_count": null, "total_count": null, "cost_usd": 1.17, "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 v1.18.1", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/d0d6404d-a144-4a7d-85db-8e1b50e071a8", "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-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "gaia", "name": "GAIA"}}, {"success_rate": 64.7, "resolved_count": null, "total_count": null, "cost_usd": 1.5992, "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-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "wildclawbench", "name": "WildClawBench"}}, {"success_rate": 62.5, "resolved_count": null, "total_count": null, "cost_usd": 7.83, "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 v1.18.1", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/95f5d2b1-910c-403c-bb17-1b10a033470e", "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-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "commit0", "name": "Commit0"}}, {"success_rate": 50.0, "resolved_count": null, "total_count": null, "cost_usd": 1.81, "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 v1.18.1", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/472e594d-c161-4836-8a93-474c2e1e4168", "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-4-8-opus", "name": "Claude 4.8 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "swe-bench-multimodal", "name": "SWE-bench Multimodal (OpenHands Index)"}}], "harnesses": [{"slug": "claude-code", "name": "Claude Code", "vendor": "Anthropic", "repo_url": "https://github.com/anthropics/claude-code", "openrouter_icon_url": "https://claude.ai/favicon.ico"}, {"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"}]}