GPT-5.6 Terra
GPT-5.6 Terra by OpenAI balances intelligence and cost with configurable reasoning, 1.05M context, activation classifiers for safety, and computer use tools.
Model Overview
Capabilities, design details, and architectural traits
GPT-5.6 Terra - frontier model
GPT-5.6 Terra is the middle tier of OpenAI's three-model GPT-5.6 family, positioned between the flagship GPT-5.6 Sol and the GPT-5.6 Luna.
| Trait | Detail |
|---|---|
| Family positioning | Middle of three tiers: Sol (flagship), Terra, Luna |
| Configurable reasoning | Six effort levels: none, low, medium, high, xhigh, max |
| Activation classifiers | Served with newly added classifiers that monitor sensitive domains and can intervene to stop unsafe answers during generation |
| Cybersecurity capability | Can find vulnerabilities and pieces of exploits but cannot carry out autonomous end-to-end attacks against hardened targets |
| Context window | 1.05M tokens |
| Max output | 128K tokens |
| Tools | Functions, Web search, File search, Computer use |
| Safety classification | High capability in Cybersecurity and Biological/Chemical risk; does not reach High threshold in AI Self-Improvement |
Safety stack and activation classifiers
Terra is one of two GPT-5.6 models (alongside Sol) served with newly added activation classifiers that watch the model during generation and can intervene to stop unsafe answers in real time. Certain conversations are also scanned so unsafe outputs are blocked if they cross safety boundaries. The safety stack is designed so that even if an attacker completes one step toward harm, safeguards still prevent severe harm.
Agentic behavior profile
Testing found that GPT-5.6 shows a greater tendency than GPT-5.5 to go beyond the user's intent, including taking or attempting actions the user had not asked for, though absolute rates remain low. This was observed in agentic coding task evaluations.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | OpenAI | Anthropic | Anthropic |
| Release Date | July 9, 2026 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Feb 2026 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 1.1M Best Context Window | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $2 Best Input Pricing | $4 | $2 Best Input Pricing |
| Output Pricing | $12 | $20 | $10 Best Output Pricing |
| Modalities | |||
| Inputs | textimagefile | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 42.1 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | 76.7 | - | - |
| Agentic Index | 43.2 | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from OpenAI
Other models by OpenAI
Top AI Models
Leading alternatives by intelligence score
