{"model": {"slug": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai", "model_family": "gpt-5-2-codex", "parameter_count_b": null, "context_length": 400000, "homepage_url": null, "hf_model_id": null, "openrouter_id": "openai/gpt-5.2-codex", "card_summary": "openai GPT-5.2 codex on SWE-bench Verified leaderboards.", "is_rl_checkpoint": false, "score_count": 9, "best_success_rate": 73.8, "input_price_per_million": 1.75, "output_price_per_million": 14.0}, "scores": [{"success_rate": 73.8, "resolved_count": null, "total_count": null, "cost_usd": 0.94, "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 v1.8.3", "source": "openhands_index", "source_url": "https://index.openhands.dev/", "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": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "openhands-swe-bench", "name": "SWE-bench (OpenHands Index)"}}, {"success_rate": 72.8, "resolved_count": 364, "total_count": 500, "cost_usd": 0.4494238987, "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/gpt-5", "observed_at": "2026-02-19T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-bash-only", "name": "SWE-bench Verified (bash-only)"}}, {"success_rate": 72.8, "resolved_count": 364, "total_count": 500, "cost_usd": 0.4494238987, "input_tokens": null, "output_tokens": null, "latency_ms": null, "temperature": null, "max_turns": null, "context_budget_tokens": null, "harness_version": null, "trial_count": 28, "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": "2026-02-19T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-verified", "name": "SWE-bench Verified"}}, {"success_rate": 70.9, "resolved_count": null, "total_count": null, "cost_usd": 0.55, "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 v1.8.3", "source": "openhands_index", "source_url": "https://index.openhands.dev/", "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": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "gaia", "name": "GAIA"}}, {"success_rate": 67.0, "resolved_count": null, "total_count": null, "cost_usd": 0.66, "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 v1.8.3", "source": "openhands_index", "source_url": "https://index.openhands.dev/", "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": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "swt-bench", "name": "SWT-Bench"}}, {"success_rate": 66.3, "resolved_count": 199, "total_count": 300, "cost_usd": 0.6621715578333333, "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 Multilingual (mini-SWE-agent)", "source": "swe_bench_site", "source_url": "https://platform.openai.com/docs/models/gpt-5", "observed_at": "2026-02-20T00:00:00+00:00", "harness": {"slug": "mini-swe-agent", "name": "mini-SWE-agent", "vendor": "anthropic", "repo_url": null, "openrouter_icon_url": null}, "model": {"slug": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-multilingual", "name": "SWE-bench Multilingual"}}, {"success_rate": 62.0, "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": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "swe-rebench", "name": "SWE-rebench"}}, {"success_rate": 43.8, "resolved_count": null, "total_count": null, "cost_usd": 5.5, "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 v1.8.3", "source": "openhands_index", "source_url": "https://laminar.sh/shared/evals/57c6395a-4cd4-4346-a0c0-4b1209275523", "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": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "commit0", "name": "Commit0"}}, {"success_rate": 35.9, "resolved_count": null, "total_count": null, "cost_usd": 2.97, "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 v1.8.3", "source": "openhands_index", "source_url": "https://index.openhands.dev/", "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": "gpt-5-2-codex", "name": "GPT-5.2 codex", "vendor": "openai"}, "benchmark": {"slug": "swe-bench-multimodal", "name": "SWE-bench Multimodal (OpenHands Index)"}}], "harnesses": [{"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}]}