{"benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)", "description": "Scientific programming benchmark (SciCode) on HAL.", "category": "eval", "split": null, "language": null, "task_count": 80, "homepage_url": "https://hal.cs.princeton.edu/scicode", "paper_url": "https://arxiv.org/abs/2510.11977", "hf_dataset_id": null, "official_leaderboard_url": "https://hal.cs.princeton.edu/scicode", "is_public_leaderboard": true, "harness_count": 3, "model_count": 15, "score_count": 30, "best_success_rate": 9.23}, "protocol": {"eval_protocol": "HAL SciCode (HAL) accuracy", "temperature": null, "max_turns": null, "context_budget_tokens": null, "n": 30}, "rankings": {"benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)", "description": "Scientific programming benchmark (SciCode) on HAL.", "category": "eval", "split": null, "language": null, "task_count": 80, "homepage_url": "https://hal.cs.princeton.edu/scicode", "paper_url": "https://arxiv.org/abs/2510.11977", "hf_dataset_id": null, "official_leaderboard_url": "https://hal.cs.princeton.edu/scicode", "is_public_leaderboard": true, "harness_count": null, "model_count": null, "score_count": null, "best_success_rate": null}, "sort": "success_rate", "rows": [{"rank": 1, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "o3", "name": "o3", "vendor": "openai"}, "success_rate": 9.23, "resolved_count": null, "total_count": null, "cost_usd": 1.3889, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 2, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-4-1-opus", "name": "Claude 4.1 Opus", "vendor": "Anthropic"}, "success_rate": 7.69, "resolved_count": null, "total_count": null, "cost_usd": 7.8141, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 3, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-4-1-opus-high", "name": "Claude 4.1 Opus (high)", "vendor": "Anthropic"}, "success_rate": 6.92, "resolved_count": null, "total_count": null, "cost_usd": 6.8817, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 4, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "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, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 5, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gpt-4-1", "name": "GPT-4.1", "vendor": "openai"}, "success_rate": 6.15, "resolved_count": null, "total_count": null, "cost_usd": 0.0352, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 6, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gpt-5-medium", "name": "GPT-5 (medium)", "vendor": "OpenAI"}, "success_rate": 6.15, "resolved_count": null, "total_count": null, "cost_usd": 2.419, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 7, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "o3", "name": "o3", "vendor": "openai"}, "success_rate": 4.62, "resolved_count": null, "total_count": null, "cost_usd": 0.0754, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 8, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "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, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 9, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-3-7-sonnet-high", "name": "Claude 3.7 Sonnet (high)", "vendor": "Anthropic"}, "success_rate": 4.62, "resolved_count": null, "total_count": null, "cost_usd": 2.5546, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 10, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "deepseek-v3-march-2025", "name": "DeepSeek V3 (March 2025)", "vendor": "DeepSeek"}, "success_rate": 3.08, "resolved_count": null, "total_count": null, "cost_usd": 0.0099, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 11, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-3-7-sonnet-high", "name": "Claude 3.7 Sonnet (high)", "vendor": "Anthropic"}, "success_rate": 3.08, "resolved_count": null, "total_count": null, "cost_usd": 0.0624, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 12, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-3-7-sonnet", "name": "Claude 3.7 Sonnet", "vendor": "Anthropic"}, "success_rate": 3.08, "resolved_count": null, "total_count": null, "cost_usd": 0.755, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 13, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "o3", "name": "o3", "vendor": "openai"}, "success_rate": 3.08, "resolved_count": null, "total_count": null, "cost_usd": 0.8373, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 14, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-4-5-sonnet", "name": "Claude 4.5 Sonnet", "vendor": "Anthropic"}, "success_rate": 3.08, "resolved_count": null, "total_count": null, "cost_usd": 1.3036, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 15, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-3-7-sonnet-high", "name": "Claude 3.7 Sonnet (high)", "vendor": "Anthropic"}, "success_rate": 3.08, "resolved_count": null, "total_count": null, "cost_usd": 2.3519, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 16, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-3-7-sonnet", "name": "Claude 3.7 Sonnet", "vendor": "Anthropic"}, "success_rate": 3.08, "resolved_count": null, "total_count": null, "cost_usd": 2.3926, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 17, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "o4-mini", "name": "o4 mini", "vendor": "openai"}, "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, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 18, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gemini-2-0-flash", "name": "Gemini 2.0 Flash", "vendor": "Google"}, "success_rate": 1.54, "resolved_count": null, "total_count": null, "cost_usd": 0.0015, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 19, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gemini-2-0-flash", "name": "Gemini 2.0 Flash", "vendor": "Google"}, "success_rate": 1.54, "resolved_count": null, "total_count": null, "cost_usd": 0.0654, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 20, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gpt-4-1", "name": "GPT-4.1", "vendor": "openai"}, "success_rate": 1.54, "resolved_count": null, "total_count": null, "cost_usd": 0.8674, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 21, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gpt-4-1", "name": "GPT-4.1", "vendor": "openai"}, "success_rate": 1.54, "resolved_count": null, "total_count": null, "cost_usd": 0.9234, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 22, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-4-5-sonnet-high", "name": "Claude 4.5 Sonnet (high)", "vendor": "Anthropic"}, "success_rate": 1.54, "resolved_count": null, "total_count": null, "cost_usd": 1.4768, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 23, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "deepseek-r1-may-2025", "name": "DeepSeek R1 (May 2025)", "vendor": "DeepSeek"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 0.0274, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 24, "harness": {"slug": "scicode-zero-shot-agent", "name": "Scicode Zero Shot Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-3-7-sonnet", "name": "Claude 3.7 Sonnet", "vendor": "Anthropic"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 0.0638, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 25, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "deepseek-v3-march-2025", "name": "DeepSeek V3 (March 2025)", "vendor": "DeepSeek"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 0.6514, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 26, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "deepseek-r1-may-2025", "name": "DeepSeek R1 (May 2025)", "vendor": "DeepSeek"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 0.7202, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 27, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "gemini-2-0-flash", "name": "Gemini 2.0 Flash", "vendor": "Google"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 0.7686, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 28, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "deepseek-v3-march-2025", "name": "DeepSeek V3 (March 2025)", "vendor": "DeepSeek"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 2.742, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 29, "harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "scaffold_code": null, "repo_url": null, "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "claude-4-5-haiku", "name": "Claude 4.5 Haiku", "vendor": "Anthropic"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 2.9045, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}, {"rank": 30, "harness": {"slug": "hal-generalist-agent", "name": "HAL Generalist Agent", "vendor": "Princeton PLI", "scaffold_code": null, "repo_url": "https://github.com/princeton-pli/hal-harness", "homepage_url": "https://hal.cs.princeton.edu/", "openrouter_icon_url": null, "leaderboard_icon_url": null}, "model": {"slug": "deepseek-r1-january-2025", "name": "DeepSeek R1 (January 2025)", "vendor": "DeepSeek"}, "success_rate": 0.0, "resolved_count": null, "total_count": null, "cost_usd": 6.0847, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "is_harnessrl_measured": false, "eval_protocol": "HAL SciCode (HAL) accuracy", "source": "hal_site", "source_url": "https://hal.cs.princeton.edu/scicode", "observed_at": null}]}}