GPT-5.2
Explore GPT-5.2’s published benchmark results and provider offerings below. More comparable task results are needed to describe its strongest areas.
Updated: Data as of:
Strongest results
| Benchmark | Measures | Score |
|---|---|---|
| SWE-bench ↗Verified / mini-SWE-agent 2.0.0 / Bash Only | Coding agents | 72.8%#4 of 7 |
| Humanity's Last Exam ↗Stable HLE — text-only subset | Expert reasoning | 37.7%#39 of 68 |
Model details
| Model | GPT-5.2 |
|---|---|
| Agent Index v1.2 | Insufficient coverage (0/7) · Preview |
| Creator | OpenAI (opens in a new tab) |
| Availability | Proprietary |
| Context size | Not reported |
| Max output | Not reported |
| Knowledge cutoff | Not reported |
| License | Not reported |
Pricing and performance
Per provider and reported configuration. Prices are USD per 1M tokens.
No provider pricing or performance reports have been added for this model yet.
Reported benchmarks
All published results in our dataset, with versions, evaluation settings, and sources.
| Benchmark | Score | Evaluation setup | Source |
|---|---|---|---|
| Humanity's Last ExamStable HLE — full setAccuracy · Higher is better | 45.5% | Published comparison table; DeepSeek dagger denotes text-only. Other rows use full text and image set. · subset: full · Tools enabled | Moonshot AI ↗ (opens in a new tab)provider reported |
| Humanity's Last ExamStable HLE — full setAccuracy · Higher is better | 34.5% | Published comparison table; DeepSeek dagger denotes text-only. Other rows use full text and image set. · subset: full · No tools | Moonshot AI ↗ (opens in a new tab)provider reported |
| Humanity's Last ExamStable HLE — text-only subsetAccuracy · Higher is better | 37.7% | question count: 2158 · reported precision decimals: 1 · GPT-5.2 (xhigh); original HLE; 2,158 text-only questions; no tools; pass@1; chart labels rounded to one decimal. · source model href: /models/gpt-5-2 · subset: text · No tools | Artificial Analysis (opens in a new tab)independent |
| LifeValNovember 2026Unweighted average reward · Higher is better | 49.3% | aggregation: Average attempts within each task, then average all 169 task means equally. No willingness-to-pay weighting. Reward multiplied by 100; published display values rounded to one decimal place. · Effort not specified in chart · tasks: 169 | OpenAI ↗ (opens in a new tab)benchmark publisher · 2026-11-02 |
| SWE-benchVerified / mini-SWE-agent 2.0.0 / Bash OnlyResolved · Higher is better | 72.8% | attempts: 1 · harness: mini-SWE-agent · harness version: 2.0.0 · reasoning effort: high · Official Bash Only leaderboard; mini-SWE-agent 2.0.0; 500 tasks; one attempt; effort high · tasks: 500 | SWE-bench ↗ (opens in a new tab)benchmark publisher · 2026-02-17 |
| τ²-benchStandard · OverallPass@1 · Higher is better | 61.6% | benchmark: tau2 · domains: retail, airline, telecom · metric: pass^1 · reasoning: none · selection: Standard checked; Custom and Legacy unchecked; Overall tab. Do not merge with banking, voice, legacy or custom results. · submitting organization: Sierra · track: Standard · user simulator: gpt-5.2 | Sierra ↗ (opens in a new tab)benchmark publisher |
| τ²-benchStandard · OverallPass@1 · Higher is better | 84.8% | benchmark: tau2 · domains: retail, airline, telecom · metric: pass^1 · reasoning: high · selection: Standard checked; Custom and Legacy unchecked; Overall tab. Do not merge with banking, voice, legacy or custom results. · submitting organization: Sierra · track: Standard · user simulator: gpt-5.2 | Sierra ↗ (opens in a new tab)benchmark publisher |