GPT-5.2
Multimodal
#1GPQA Diamond
#1MMMU-Pro with Tools
#1GDPVal
+8 more
by OpenAI
+
+
+
+
About
GPT-5.2, released by OpenAI on December 11, 2025, is a large language model from the GPT-5 family that improves on GPT-5 in general intelligence, long-context understanding, agentic tool-calling, and vision. It features a 400K token context window, 128K maximum output tokens, and a knowledge cutoff of August 2025. GPT-5.2 targets long-context coding tasks, extended document analysis, and complex agentic workflows requiring reliable instruction following.
+
+
+
+
Timeline
ReleasedDec 11, 2025
Knowledge CutoffAug 1, 2025
+
+
+
+
Specifications
Capabilities
Multimodal
+
+
+
+
License & Family
License
Proprietary
Performance Overview
Performance metrics and category breakdown
Overall Performance
18 benchmarks
Average Score
67.7%
Best Score
98.7%
High Performers (80%+)
6Top Categories
Science
93.2%
Knowledge
89.6%
Multimodal
80.0%
Agents
69.9%
Coding
65.2%
+
+
+
+
All Benchmark Results for GPT-5.2
Complete list of benchmark scores with detailed information
| TAU2-Bench Telecom | Agents | 98.70 | 98.7% | Self-reported | |
| GPQA Diamond | Science | 93.20 | 93.2% | Unverified | |
| MMMLU | Knowledge | 89.60 | 89.6% | Self-reported | |
| TAU2-Bench Retail | Agents | 82.00 | 82.0% | Self-reported | |
| MMMU-Pro with Tools | Multimodal | 80.40 | 80.4% | Self-reported | |
| SWE Bench Verified | Coding | 80.00 | 80.0% | Unverified | |
| MMMU-Pro | Multimodal | 79.50 | 79.5% | Self-reported | |
| BrowseComp | Agents | 77.90 | 77.9% | Self-reported | |
| GDPVal AA ELO | Agents | 1462.00 | 73.1% | Self-reported | |
| Terminal Bench 2.0 | Coding | 64.70 | 64.7% | Unverified |
Showing 1 to 10 of 18 benchmarks