GPT-5.6 Sol
currentGPT-5.6 Sol is OpenAI's highest-capability GPT-5.6 model, built for difficult reasoning, professional work, coding agents, and long-running tool use. It is the family choice when result quality matters more than latency or token cost; Terra is the more economical default for many production workloads.
Context window
1.05M tokens
Maximum output
128K tokens
Input price
$5 / 1M
Output price
$30 / 1M
Best for
- The hardest coding and software-agent tasks
- Complex professional research and knowledge work
- Long-running agents that must plan, use tools, and recover
- Large document or codebase analysis up to 1.05 million tokens
Not ideal for
- High-volume work where Terra or Luna is sufficient
- Latency-sensitive chat and simple extraction
- Native audio or video generation
- Fine-tuning or self-hosting
PromptHQ 100-word challenge
We give every model the same miniature science-fiction mystery. The response is short enough to read in one chat bubble, while the premise tests creative writing, compression, foreshadowing, causal coherence, and instruction following.
GPT-5.6 Sol100 wordsWrite a science-fiction mystery in exactly 100 words.
A three-person research station on Europa receives a message from itself, timestamped 11 minutes in the future:
"Do not open the greenhouse door."
The door is opened anyway.
Plant two subtle clues. Give the warning a logical explanation without aliens, time travel, hallucinations, or simulations. The final sentence must change what the reader thought the warning meant.
Output only the story. Do not explain it.
At 03:14, Europa Station received its emergency packet, stamped 03:25: “Do not open the greenhouse door.”
Mara found no transmission in the comm buffer. Ilyan noticed frost feathers pointing inward around the window seal, though both pressure gauges agreed. Chen opened the door anyway.
Air shrieked into the greenhouse. A cracked sampling tube had depressurized it, while ice pinned its gauge. The greenhouse controller had issued the warning; its radiation-damaged clock ran eleven minutes fast, and the local network labeled every internal sender “Europa Station.”
Later, seedlings blackened wherever their breath touched them. The quarantine had been protecting the plants.
Run the same prompt yourself
Open GPT-5.6 Sol in Multichat with the evaluation prompt prefilled.
What the response revealed
The response contains exactly 100 whitespace-delimited words.
It outputs only the story and respects the exclusions and required premise.
A fast damaged controller clock, internal sender labeling, a cracked tube, and a pinned gauge explain both the message and the danger.
Inward frost and falsely agreeing gauges subtly prepare the pressure-loss explanation.
The final sentence convincingly changes the warning from crew protection to plant quarantine.
The prose is economical, concrete, and complete without feeling like a compressed synopsis.
The model receives the prompt without web access or external tools. GPT-5.6 Sol is run through Codex at medium reasoning effort. We preserve the response as generated apart from display rendering.
Performance
GPT-5.6 Sol benchmarks
Benchmark scores are sensitive to reasoning effort, harness, tools, token budget, prompt format, sampling, and evaluation date. Scores here retain their source and should not be treated as directly interchangeable unless the underlying setup matches.
Agents' Last Exam
52.7%
professional agent tasks · OpenAI2
GDPval-AA v2
1,747.8 Elo
professional knowledge work · OpenAI2
SWE-Bench Pro
64.6%
software engineering · OpenAI2
Terminal-Bench 2.1
88.8%
terminal and agentic coding · OpenAI2
MMMU Pro, no tools
83%
multimodal reasoning · OpenAI2
OpenAI MRCR v2, 512K-1M
73.8%
long-context retrieval · OpenAI2
Family position
GPT-5.6 positioning
Sol is the flagship of the GPT-5.6 family. Terra trades some peak capability for half-price tokens, while Luna targets speed and volume.
GPT-5.6 Sol
This modelFlagship capability
$5 input
$30 output
GPT-5.6 Terra
Balanced intelligence and cost
$2.50 input
$15 output
GPT-5.6 Luna
Fast, cost-sensitive workloads
$1 input
$6 output
API pricing
Per million text tokens
Input
$5
Cached input
$0.50
Cache write
$6.25
Output
$30
Where GPT-5.6 Sol stands out
Built for demanding work
The hardest coding and software-agent tasks are central to the model's positioning, rather than an incidental capability.
Long-context capacity
The published context window is 1,050,000 tokens, making the model a candidate for large documents, repositories, and sustained agent state.
Reasoning and tools
Reasoning is supported with none, low, medium, high, xhigh, max provider setting(s), and the model can participate in tool-using workflows through its available API surface.
Limitations to know
Benchmarks are configuration-sensitive
Scores can move substantially with the harness, tool access, effort setting, token budget, and evaluator. Treat the table as evidence, not a universal ranking.
Context size is not guaranteed recall
A large advertised window does not mean every detail is retrieved reliably at maximum length. Validate representative long-context workloads before deployment.
Product access differs from model capability
PromptHQ and gateway limits may expose fewer modalities, tools, or tokens than the provider's first-party API.
Capabilities and specifications
Knowledge cutoff
Feb 16, 2026
Inputs
text, image
Reasoning
Supported
Default effort
medium
Supported API features
Frequently Asked Questions
Sources
- 1GPT-5.6 Sol model
OpenAI · official documentation
- 2GPT-5.6: Frontier intelligence that scales with your ambition
OpenAI · official launch post
- 3PromptHQ model registry
PromptHQ · internal product configuration