GPT-5.5
Parameters
Undisclosed
Context Length
1M
Modality
Multimodal
Architecture
Undisclosed
License
Proprietary
Release Date
23 Apr 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Input: $5.00 · Output: $30.00
Evaluation Benchmarks
Rank
#19
| Benchmark | Score | Rank |
|---|---|---|
Data Analysis | max 0.82 | 🥉 3 |
StackUnseen | 0.91 | 5 |
Graduate-Level QA | 0.936 | 5 |
Coding | max 0.82 | 8 |
General | max 0.80 | 9 |
LiveBench Average | max 0.80 | 9 |
Mathematics | max 0.96 | 10 |
Reasoning | max 0.90 | 13 |
Agent Arena | max 0.04 Standard 0.01 | 16 30 |
General Text | high 1481 Standard 1477 | 20 25 |
Coding Index | max 0.75 high 0.72 medium 0.71 low 0.61 Standard 0.56 | 21 35 36 65 72 |
Agentic Coding | max 0.54 | 28 |
Web Development | max 1512 high 1487 Standard 1455 | 48 54 59 |
Agentic Index | max 0.36 high 0.34 medium 0.29 | 50 54 62 |
Intelligence Index | max 0.38 high 0.37 medium 0.34 low 0.31 Standard 0.23 | 59 64 76 91 141 |
Rankings
Overall Rank
#19
Coding Rank
#20
About GPT-5.5
GPT-5.5 is OpenAI's most capable and efficient frontier model for agentic work, delivering state-of-the-art performance on Terminal-Bench 2.0 (82.7%), SWE-Bench Pro (58.6%), OSWorld-Verified (78.7%), BrowseComp (84.4%), ARC-AGI-2 (85.0%), and FrontierMath Tier 4 (35.4%). Matches GPT-5.4 per-token latency in production while using significantly fewer tokens. Supports text, image, and tool inputs with a 1M context window. Released April 23, 2026 in ChatGPT and Codex; API available as gpt-5.5 at $5/M input and $30/M output tokens.
Technical Specifications
Architecture specifications are undisclosed for proprietary models.
Resources
About GPT-5.5
GPT-5.5 is OpenAI's smartest and most intuitive model yet, and the next step toward a new way of getting work done on a computer. It excels at agentic coding, computer use, knowledge work, and scientific research — areas where progress depends on reasoning across context and taking action over time. It matches GPT-5.4 per-token latency in real-world serving while performing at a significantly higher level of intelligence, and uses fewer tokens to complete the same tasks. State-of-the-art on Terminal-Bench 2.0 (82.7%), SWE-Bench Pro (58.6%), OSWorld-Verified (78.7%), and ARC-AGI-2 (85.0%). Features a 1M context window and costs $5/M input tokens and $30/M output tokens. Released April 23, 2026.