GPT-5.6 Terra
GPTBalanced
ThinkingTool UseVisionStructured Output
About this model
GPT-5.6 Terra is a balanced model in OpenAI's GPT-5.6 series, positioned between the flagship Sol tier and the cost-efficient Luna tier. It is suited for everyday coding, reasoning, and agentic...
Performance Tier
Balanced
GPT-5.6 Terra is a balanced model from GPT : strong performance at a reasonable price.
Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.
Pricing
This model is included in Elosia plansHigh
High pricing. Reserve this level for tasks that demand maximum quality.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $2.50 |
| Output (completion) | $15.00 |
| Cache read | $0.250 |
| Cache write | $3.13 |
Capabilities
Context Length1.1M
Max Output Tokens128K
TokenizerGPT
Inputfile, image, text
Outputtext
Release DateJuly 9, 2026
Benchmarks
General Intelligence
MMLU
Not reported
Mathematics
MATH-500
Not reported
Programming
HumanEval
Not reported
Reasoning
ARC-AGI-2
67.1%
Agentic
SWE-bench Pro
63.4%
Terminal-Bench 2.1
87.4%
Where does GPT-5.6 Terra stand?
Compare its performance index against every other model.
Recommended Use Cases
CodingAnalysisGeneral ChatResearchData Extraction
Strengths
- Balanced tier between the Sol flagship and the Luna value tier, at $2.50 / $15 per M tokens
- GPT-5.5-competitive quality at roughly half the cost, per OpenAI positioning
- Strong agentic results close to Sol: 87.4 on Terminal-Bench 2.1 and 63.4 on SWE-bench Pro
- 1M-token context window with multimodal (text, image, file) input and native tool use
Limitations
- Abstract reasoning trails Sol by a wide margin (ARC-AGI-2 67.1 vs 85.4 at high effort)
- Only agentic benchmarks published; no MMLU, GPQA or math scores for direct comparison
- Reasoning overhead makes it slower and more token-hungry than Luna on simple tasks
Frequently asked questions
Resources
This model may use your data for training
