GPT-5.2
OpenAI
OpenAI's late-2025 flagship (December 2025) with a massive 400K context window and enhanced reasoning capabilities, now superseded by the GPT-5.4 and GPT-5.5 generation (April 2026). Available in three tiers (Instant, Thinking, Pro) for different use cases. First model to exceed 90% on ARC-AGI benchmark, representing a significant leap in general intelligence. Excels at complex reasoning, coding, and agentic workflows with hybrid reasoning that balances speed and depth.
Strengths
- Massive 400K token context window - largest among major models (3x larger than GPT-4)
- 128K max output tokens enables generation of full documents and codebases
- Exceptional benchmark performance: 93.2% GPQA Diamond, 100% AIME, 70.9% GDPval
- Three model tiers (Instant/Thinking/Pro) optimize for speed vs reasoning depth
- First model above 90% on ARC-AGI benchmark (significant AGI progress milestone)
- Strong multimodal capabilities with native audio and vision understanding
- Hybrid reasoning with configurable effort for optimal speed-accuracy tradeoffs
Caveats
- 40% more expensive than GPT-5 ($1.75 vs $1.25 input, $14 vs $10 output)
- Thinking tier has higher latency due to extended reasoning process
- Knowledge cutoff August 2025 (older than some competitors with more recent training)
- Requires careful tier selection - Pro tier significantly more expensive for marginal gains
- Very large context windows may have higher latency for complex queries
- Vendor
- OpenAI
- Context Window
- 400,000 tokens
- Max Output
- 128,000 tokens
- License
- Proprietary
- Release Date
- 2025-12-11
- Training Cutoff
- 2025-08-31
- Modalities
- textimageaudio
| Benchmark | Score |
|---|---|
| GPQA Diamond | 93.2 |
| AIME 2025 | 100.0 |
| GDPval | 70.9 |
| SWE-Bench Pro | 55.6 |
Capabilities
Vision
Audio
Video
Tool Use
Resources
Pricing
$1.75 per 1M input tokens, $14.00 per 1M output tokens (40% more than GPT-5, reflecting advanced capabilities)