{"model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai", "model_family": "o4", "parameter_count_b": null, "context_length": 200000, "homepage_url": null, "hf_model_id": null, "openrouter_id": "openai/o4-mini", "card_summary": "openai o4 mini on SWE-bench Verified leaderboards.", "is_rl_checkpoint": false, "score_count": 23, "best_success_rate": 84.45, "input_price_per_million": 1.1, "output_price_per_million": 4.4}, "scores": [{"success_rate": 84.45, "resolved_count": null, "total_count": 1055, "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": "LiveCodeBench generation pass@1 mean", "source": "livecodebench", "source_url": "https://livecodebench.github.io/leaderboard.html", "observed_at": null, "harness": {"slug": "livecodebench", "name": "Livecodebench", "vendor": "LiveCodeBench", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "livecodebench-generation", "name": "LiveCodeBench (code generation)"}}, {"success_rate": 64.6, "resolved_count": 323, "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.swebench.com/verified", "observed_at": "2025-05-03T00:00:00+00:00", "harness": {"slug": "patchpilot", "name": "Patchpilot", "vendor": "Ucsb Mlsec", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 54.55, "resolved_count": null, "total_count": null, "cost_usd": 0.3599, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 165, "is_harnessrl_measured": false, "eval_protocol": "HAL GAIA (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/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": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-gaia", "name": "GAIA (HAL)"}}, {"success_rate": 53.09, "resolved_count": null, "total_count": null, "cost_usd": 0.1822, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 135, "is_harnessrl_measured": false, "eval_protocol": "HAL USACO (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/usaco", "observed_at": null, "harness": {"slug": "usaco-episodic-semantic", "name": "USACO Episodic + Semantic", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-usaco", "name": "USACO (HAL)"}}, {"success_rate": 50.0, "resolved_count": null, "total_count": null, "cost_usd": 5.0706, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 49, "is_harnessrl_measured": false, "eval_protocol": "HAL SWE-bench Verified Mini (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/swebench_verified_mini", "observed_at": null, "harness": {"slug": "swe-agent", "name": "SWE-agent", "vendor": "SWE-agent", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-swe-bench-verified-mini", "name": "SWE-bench Verified Mini (HAL)"}}, {"success_rate": 47.88, "resolved_count": null, "total_count": null, "cost_usd": 0.4897, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 165, "is_harnessrl_measured": false, "eval_protocol": "HAL GAIA (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/gaia", "observed_at": null, "harness": {"slug": "hf-open-deep-research", "name": "HF Open Deep Research", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-gaia", "name": "GAIA (HAL)"}}, {"success_rate": 45.0, "resolved_count": 225, "total_count": 500, "cost_usd": 0.2099733328, "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://platform.openai.com/docs/models/o4-mini", "observed_at": "2025-07-26T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-bash-only", "name": "SWE-bench Verified (bash-only)"}}, {"success_rate": 45.0, "resolved_count": 225, "total_count": 500, "cost_usd": 0.2099733328, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 23, "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": "2025-07-26T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 43.64, "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": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "swe-rebench", "name": "SWE-rebench"}}, {"success_rate": 36.0, "resolved_count": null, "total_count": null, "cost_usd": 0.1428, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 50, "is_harnessrl_measured": false, "eval_protocol": "HAL TAU-bench Airline (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/taubench_airline", "observed_at": null, "harness": {"slug": "tau-bench-tool-calling", "name": "TAU-bench Tool Calling", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-tau-bench-airline", "name": "TAU-bench Airline (HAL)"}}, {"success_rate": 33.85, "resolved_count": 175, "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://sites.google.com/view/guirepair", "observed_at": "2025-05-31T00:00:00+00:00", "harness": {"slug": "guirepair", "name": "GUIRepair", "vendor": "GUIRepair", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-multimodal-site", "name": "SWE-bench Multimodal (site)"}}, {"success_rate": 31.67, "resolved_count": null, "total_count": null, "cost_usd": 0.5412, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 300, "is_harnessrl_measured": false, "eval_protocol": "HAL Online Mind2Web (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/online_mind2web", "observed_at": null, "harness": {"slug": "seeact", "name": "SeeAct", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-online-mind2web", "name": "Online Mind2Web (HAL)"}}, {"success_rate": 27.45, "resolved_count": null, "total_count": null, "cost_usd": 0.1096, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 102, "is_harnessrl_measured": false, "eval_protocol": "HAL ScienceAgentBench (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scienceagentbench", "observed_at": null, "harness": {"slug": "sab-self-debug", "name": "SAB Self Debug", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-scienceagentbench", "name": "ScienceAgentBench (HAL)"}}, {"success_rate": 23.84, "resolved_count": null, "total_count": null, "cost_usd": 0.5463, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 30, "is_harnessrl_measured": false, "eval_protocol": "HAL AssistantBench (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/assistantbench", "observed_at": null, "harness": {"slug": "browser-use", "name": "Browser-Use", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-assistantbench", "name": "AssistantBench (HAL)"}}, {"success_rate": 19.61, "resolved_count": null, "total_count": null, "cost_usd": 0.758, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 102, "is_harnessrl_measured": false, "eval_protocol": "HAL ScienceAgentBench (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scienceagentbench", "observed_at": null, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-scienceagentbench", "name": "ScienceAgentBench (HAL)"}}, {"success_rate": 18.33, "resolved_count": null, "total_count": null, "cost_usd": 0.6715, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 300, "is_harnessrl_measured": false, "eval_protocol": "HAL Online Mind2Web (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/online_mind2web", "observed_at": null, "harness": {"slug": "browser-use", "name": "Browser-Use", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-online-mind2web", "name": "Online Mind2Web (HAL)"}}, {"success_rate": 18.0, "resolved_count": null, "total_count": null, "cost_usd": 0.4114, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 50, "is_harnessrl_measured": false, "eval_protocol": "HAL TAU-bench Airline (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/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": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-tau-bench-airline", "name": "TAU-bench Airline (HAL)"}}, {"success_rate": 17.78, "resolved_count": null, "total_count": null, "cost_usd": 0.7064, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 45, "is_harnessrl_measured": false, "eval_protocol": "HAL CORE-Bench Hard (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/corebench_hard", "observed_at": null, "harness": {"slug": "core-agent", "name": "CORE-Agent", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-corebench-hard", "name": "CORE-Bench Hard (HAL)"}}, {"success_rate": 15.56, "resolved_count": null, "total_count": null, "cost_usd": 0.5, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 45, "is_harnessrl_measured": false, "eval_protocol": "HAL CORE-Bench Hard (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/corebench_hard", "observed_at": null, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-corebench-hard", "name": "CORE-Bench Hard (HAL)"}}, {"success_rate": 6.15, "resolved_count": null, "total_count": null, "cost_usd": 0.0671, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 80, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"success_rate": 4.62, "resolved_count": null, "total_count": null, "cost_usd": 0.8275, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 80, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"success_rate": 2.0, "resolved_count": null, "total_count": null, "cost_usd": 0.6535, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 49, "is_harnessrl_measured": false, "eval_protocol": "HAL SWE-bench Verified Mini (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/swebench_verified_mini", "observed_at": null, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-swe-bench-verified-mini", "name": "SWE-bench Verified Mini (HAL)"}}, {"success_rate": 1.54, "resolved_count": null, "total_count": null, "cost_usd": 1.1512, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 80, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}], "harnesses": [{"slug": "browser-use", "name": "Browser-Use", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, {"slug": "core-agent", "name": "CORE-Agent", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, {"slug": "guirepair", "name": "GUIRepair", "vendor": "GUIRepair", "repo_url": null, "openrouter_icon_url": null}, {"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": "hf-open-deep-research", "name": "HF Open Deep Research", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, {"slug": "livecodebench", "name": "Livecodebench", "vendor": "LiveCodeBench", "repo_url": "https://github.com/LiveCodeBench/LiveCodeBench", "openrouter_icon_url": null}, {"slug": "patchpilot", "name": "Patchpilot", "vendor": "Ucsb Mlsec", "repo_url": "https://github.com/ucsb-mlsec/Co-PatcheR", "openrouter_icon_url": null}, {"slug": "sab-self-debug", "name": "SAB Self Debug", "vendor": null, "repo_url": null, "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": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, {"slug": "seeact", "name": "SeeAct", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, {"slug": "tau-bench-tool-calling", "name": "TAU-bench Tool Calling", "vendor": null, "repo_url": null, "openrouter_icon_url": null}, {"slug": "usaco-episodic-semantic", "name": "USACO Episodic + Semantic", "vendor": null, "repo_url": null, "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}]}