GPT-5 Mini
OpenAI·Released Aug 7, 2025
Updated Oct 8, 11:55 PM ET
The score is on the board's scale (reference average 50, about 15 points per standard deviation), not a percentage.
By domain
Each domain is scored on the same scale as the overall score. Coverage is the share of the domain's kinds of evaluation with a result; it says how much evidence there is, not how good the model is.
Every result behind the score
The result as the evaluator publishes it, and the reading it gives on the board's scale. A hard evaluation can give a high reading for a modest result; results near 95% or 5% only say "at least" or "at most".
Coding
3 results · 24.6- Terminal-Bench 4.0 (AA run)Artificial Analysis · high effortNear the floor0.0%Reading at most 34.5
- Vibe Code BenchVals AI · high effort14.2%Reading 24.7
- SciCodeArtificial Analysis · high effort39.0%Reading 19.3
Research and reasoning
5 results · 27.8- Chess PuzzlesEpoch AI · high effort30.0%Reading 46.7
- FrontierMath Tiers 1–3Epoch AI · high effort46.7%Reading 39.5
- FrontierMath Tier 4Epoch AI · high effort12.2%Reading 35.1
- Humanity's Last Exam (AA run)Artificial Analysis · high effort21.5%Reading 27.7
- Mystery Game PuzzlesEpoch AI · high effortNear the floor5.0%Reading at most -0.8
Professional work
3 results · 31.6- MedCodeVals AI · high effort43.0%Reading 44.4
- τ-Bench Banking (AA run)Artificial Analysis · high effort15.5%Reading 22.9
- τ²-Bench Telecom (AA run)Artificial Analysis · high effort68.4%Reading 17.5
Knowledge and accuracy
3 results · 32.8- IFBench (AA run)Artificial Analysis · high effort75.4%Reading 53.6
- AA-LCRArtificial Analysis · high effort72.3%Reading 37.6
- SimpleQA VerifiedEpoch AI · high effort21.6%Reading 9.3
Human preference
1 result · 24.6- Arena TextLMArena · high effort1,389.7Reading 22.6
Shown, not scored
Composite indices, saturated or older evaluations, evaluations still being checked, and vision, writing and multilingual results. They do not move the score.
- LiveCodeBenchVals AI · CodingReference only86.6%
- LiveCodeBench (AA run)Artificial Analysis · CodingReference only83.8%
- SWE-bench VerifiedEpoch AI · CodingReference only64.7%
- SWE-bench VerifiedVals AI · CodingReference only60.8%
- AIME (AA run)Artificial Analysis · Research and reasoningReference only90.7%
- OTIS Mock AIMEEpoch AI · Research and reasoningReference only86.7%
- CorpFinVals AI · Professional workReference only60.2%
- LegalBenchVals AI · Professional workReference only81.8%
- MedScribeVals AI · Professional workWatching80.6%
- TaxEvalVals AI · Professional workReference only75.2%
- GPQA DiamondEpoch AI · Knowledge and accuracyReference only75.0%
- GPQA DiamondVals AI · Knowledge and accuracyReference only80.3%
- MMLU-ProVals AI · Knowledge and accuracyReference only82.2%
- AA Intelligence IndexArtificial Analysis · Composite indicesReference only16.8
- Arena VisionLMArena · VisionReference only1,183.2
- MMMU ProVals AI · VisionReference only78.9%
- MGSMVals AI · MultilingualReference only92.6%
No result yet
Scored evaluations this model has not taken. A missing result neither adds nor subtracts.
- Terminal-Bench 4.0 · Vals AI
- LiveBench Coding · LiveBench
- Code Migration · Vals AI
- ProgramBench · Vals AI
- Vibe Code Bench 1–100 · Vals AI
- CyberBench Patch · Vals AI
- APEX-SWE · Mercor
- LiveBench Reasoning · LiveBench
- MysteryMechanism · Vals AI
- Terminal-Bench Science · Vals AI
- ProofBench · Vals AI
- Finance Agent · Vals AI
- APEX-Agents · Mercor
- Harvey Legal Agent Benchmark · Vals AI
- Legal Research Bench · Vals AI
- Tax Agent Bench · Vals AI
- EMB · Vals AI
- LiveBench Data Analysis · LiveBench
- LiveBench Language · LiveBench
- LiveBench Instruction Following · LiveBench
- BullshitBench · BullshitBench