Atlas/Models/gpt-5-2

Exact model identity

GPT-5.2

OpenAI · version status: reported

Similar model names are never merged automatically. Evaluation membership and numeric result rows only reference the exact ID gpt-5-2.

Evaluation settings

Benchmark	Work	Run / group	Scope
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-official genebench-pro-paper-v1-full-standard-xhigh	full · n=129
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-high genebench-pro-paper-v1-full-standard-high	full · n=129
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-low genebench-pro-paper-v1-full-standard-low	full · n=129
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-medium genebench-pro-paper-v1-full-standard-medium	full · n=129
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-none genebench-pro-paper-v1-full-standard-none	full · n=129
SpatialBench	SpatialBench: Can Agents Analyze Real-World Spatial Biology Data?	spatialbench-paper-v2-base spatialbench-paper-v2-base	full · n=146
SpatialBench	SpatialBench 159-evaluation repository snapshot	spatialbench-repo-159-mini-swe-agent spatialbench-repo-159-mini-swe-agent	full · n=159

Numeric results

Benchmark	Work	Run / group	Metric	Value
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-official genebench-pro-paper-v1-full-standard-xhigh	Eval-level pass rate	4.9 percent
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-high genebench-pro-paper-v1-full-standard-high	Eval-level pass rate	3.5 percent
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-low genebench-pro-paper-v1-full-standard-low	Eval-level pass rate	1.1 percent
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-medium genebench-pro-paper-v1-full-standard-medium	Eval-level pass rate	2.4 percent
GeneBench-Pro	GeneBench-Pro: Evaluating Multistage Statistical Reasoning in Genomics, Quantitative Biology, and Translational Biomedicine	genebench-pro-standard-none genebench-pro-paper-v1-full-standard-none	Eval-level pass rate	0.5 percent
SpatialBench	SpatialBench: Can Agents Analyze Real-World Spatial Biology Data?	spatialbench-paper-v2-base spatialbench-paper-v2-base	Accuracy	34.02 percent
SpatialBench	SpatialBench: Can Agents Analyze Real-World Spatial Biology Data?	spatialbench-paper-v2-base spatialbench-paper-v2-base	Steps	2.1 steps per evaluation
SpatialBench	SpatialBench: Can Agents Analyze Real-World Spatial Biology Data?	spatialbench-paper-v2-base spatialbench-paper-v2-base	Latency	89.2 seconds per evaluation
SpatialBench	SpatialBench: Can Agents Analyze Real-World Spatial Biology Data?	spatialbench-paper-v2-base spatialbench-paper-v2-base	Cost	0.037 USD per evaluation
SpatialBench	SpatialBench 159-evaluation repository snapshot	spatialbench-repo-159-mini-swe-agent spatialbench-repo-159-mini-swe-agent	Accuracy	50.1 percent
SpatialBench	SpatialBench 159-evaluation repository snapshot	spatialbench-repo-159-mini-swe-agent spatialbench-repo-159-mini-swe-agent	Cost	0.6024 USD per evaluation
SpatialBench	SpatialBench 159-evaluation repository snapshot	spatialbench-repo-159-mini-swe-agent spatialbench-repo-159-mini-swe-agent	Duration	931.76 seconds per evaluation