{"model": {"slug": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google", "model_family": "gemini-3-5-flash", "parameter_count_b": null, "context_length": 1048576, "homepage_url": null, "hf_model_id": null, "openrouter_id": "google/gemini-3.5-flash", "card_summary": "google Gemini 3.5 Flash on SWE-bench Verified leaderboards.", "is_rl_checkpoint": false, "score_count": 8, "best_success_rate": 90.25, "input_price_per_million": 1.5, "output_price_per_million": 9.0}, "scores": [{"success_rate": 90.25, "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": "HAL reliability composite (mean of 4 dimension aggregates)", "source": "hal_reliability_site", "source_url": "https://hal.cs.princeton.edu/reliability/benchmark/gaia/", "observed_at": null, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "hal-reliability-gaia", "name": "HAL Reliability \u2014 GAIA"}}, {"success_rate": 88.75, "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": "HAL reliability composite (mean of 4 dimension aggregates)", "source": "hal_reliability_site", "source_url": "https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/", "observed_at": null, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "hal-reliability-tau-bench-airline", "name": "HAL Reliability \u2014 \u03c4-bench Airline (clean)"}}, {"success_rate": 81.2, "resolved_count": null, "total_count": null, "cost_usd": 2.63, "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 v1.24.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/0e76f325-8100-4491-8a2d-37fb1ba84d62", "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": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "gaia", "name": "GAIA"}}, {"success_rate": 78.6, "resolved_count": null, "total_count": null, "cost_usd": 1.53, "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 v1.24.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/60edc213-0c2b-4e3f-9c72-ba4bfce1cb32", "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": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "openhands-swe-bench", "name": "SWE-bench (OpenHands Index)"}}, {"success_rate": 76.2, "resolved_count": null, "total_count": null, "cost_usd": 1.03, "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 v1.24.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/88a0e23e-741b-4eba-a6a6-87056fb23905", "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": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "swt-bench", "name": "SWT-Bench"}}, {"success_rate": 62.67, "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": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "swe-rebench", "name": "SWE-rebench"}}, {"success_rate": 39.7, "resolved_count": null, "total_count": null, "cost_usd": 1.99, "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 v1.24.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/20d43172-1fef-4d83-af1b-01784887e7ca", "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": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "swe-bench-multimodal", "name": "SWE-bench Multimodal (OpenHands Index)"}}, {"success_rate": 37.5, "resolved_count": null, "total_count": null, "cost_usd": 7.91, "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 v1.24.0", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/f73dcfa9-a6de-4b41-bcf6-abc1d2b0ba58", "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": "gemini-3-5-flash", "name": "Gemini 3.5 Flash", "vendor": "google"}, "benchmark": {"slug": "commit0", "name": "Commit0"}}], "harnesses": [{"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "repo_url": "https://github.com/princeton-pli/hal-harness", "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}]}