{"harness": {"slug": "scicode-tool-calling-agent", "name": "Scicode Tool Calling Agent", "vendor": null, "description": "Scicode Tool Calling Agent is tracked in this catalog as a software-engineering agent harness. Public benchmark rows use this slug on leaderboards such as SWE-bench Verified or Artificial Analysis coding-agent evals.\n\nCheck each score row's `source` and `source_url` for the authoritative snapshot. Harness-level metadata here does not invent scores \u2014 only documents what the harness is and where published results come from.", "card_summary": "Scicode Tool Calling Agent coding agent harness.", "homepage_url": "https://hal.cs.princeton.edu/", "repo_url": null, "docs_url": null, "scaffold_code": null, "harbor_agent_name": null, "api_protocol": null, "status": "validated", "supports_rl": true, "score_count": 14, "best_success_rate": 9.23, "github_stars": null, "popularity_tier": null, "hrl_score": null, "hrl_rank": null, "catalog_token_volume": null, "openrouter_icon_url": null, "leaderboard_icon_url": null, "aa_coding_index": null, "aa_mean_cost_usd": null, "aa_mean_total_tokens": null}, "scores": [{"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, "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": "o3", "name": "o3", "vendor": "openai"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "claude-4-1-opus", "name": "Claude 4.1 Opus", "vendor": "Anthropic"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "claude-4-1-opus-high", "name": "Claude 4.1 Opus (high)", "vendor": "Anthropic"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "gpt-5-medium", "name": "GPT-5 (medium)", "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": 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, "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": "claude-3-7-sonnet-high", "name": "Claude 3.7 Sonnet (high)", "vendor": "Anthropic"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "claude-4-5-sonnet", "name": "Claude 4.5 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "claude-3-7-sonnet", "name": "Claude 3.7 Sonnet", "vendor": "Anthropic"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "gemini-2-0-flash", "name": "Gemini 2.0 Flash", "vendor": "Google"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "gpt-4-1", "name": "GPT-4.1", "vendor": "openai"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "claude-4-5-sonnet-high", "name": "Claude 4.5 Sonnet (high)", "vendor": "Anthropic"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "deepseek-v3-march-2025", "name": "DeepSeek V3 (March 2025)", "vendor": "DeepSeek"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "deepseek-r1-may-2025", "name": "DeepSeek R1 (May 2025)", "vendor": "DeepSeek"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}, {"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, "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": "claude-4-5-haiku", "name": "Claude 4.5 Haiku", "vendor": "Anthropic"}, "benchmark": {"slug": "hal-scicode", "name": "SciCode (HAL)"}}], "benchmarks": [{"slug": "hal-scicode", "name": "SciCode (HAL)", "category": "eval"}]}