OpenAI GPT‑5.6 Sol Released
Thursday, July 9, 2026 · Time not confirmed

Where to Buy
Details
- Brand / Manufacturer
- OpenAI
- Product type
- AI model
- Model
- GPT-5.6 Sol
- Region Availability
- Global
Release information
GPT-5.6 Sol launched for general availability following a limited preview phase. In addition to direct first-party API endpoints, the model is available via Amazon Bedrock with matching first-party token rates.
GPT-5.6 Sol is a flagship artificial intelligence model designed for complex professional tasks, software engineering, scientific reasoning, and cybersecurity. Operating as the high-capability tier within the GPT-5.6 release generation alongside Terra and Luna, Sol delivers multi-agent execution and programmable tool orchestration.
What we know
The GPT-5.6 family introduces a structured naming system where numbers indicate the overall model generation, while Sol, Terra, and Luna represent specific capability tiers. Sol serves as the premier flagship reasoning tier in this lineup, designed for complex professional tasks, software engineering, and scientific reasoning. API requests directed to the standard gpt-5.6 model alias automatically route to Sol.
Sol features a total context window capacity of 1,050,000 tokens and supports up to 128,000 output tokens per single request. The model accepts both text and image inputs, generating text responses with a knowledge cutoff of February 16, 2026. While streaming, function calling, and structured outputs are natively supported, fine-tuning is not available.
Features and specs
Sol introduces an ultra capability setting designed for high-demand workloads. This setting coordinates multiple autonomous agents across parallel workstreams to finish complex agentic tasks faster. Developers can control reasoning depth using six adjustable effort levels: none, low, medium, high, xhigh, and max, with medium serving as the default setting.
Through Programmatic Tool Calling within the Responses API, the model writes and executes lightweight scripts to filter intermediate data automatically. This capability minimizes model round trips and saves output tokens during multi-step tool interactions. Built-in system protections combine safety safeguards trained directly into the model with real-time risk monitoring.
Performance and benchmarks
In evaluations on Agents' Last Exam, which measures long-horizon professional workflows across 55 fields, Sol achieved a top score of 53.6 at maximum reasoning effort. This result surpassed Claude Fable 5 by 13.1 points. Operating at medium reasoning effort, Sol also beat Fable 5 by 11.4 points while maintaining lower estimated operational costs.
On the Artificial Analysis Coding Agent Index, Sol scored 80 at maximum reasoning, exceeding Fable 5 by 2.8 points while consuming less than half the output tokens and completion time. The model recorded state-of-the-art results on Terminal-Bench 2.1 and DeepSWE, while completing Artificial Analysis Intelligence Index tasks in 61 percent less time.
Pricing and availability
Standard API rates for Sol are set at $4.00 per million input tokens, $0.40 per million cached input tokens, and $20.00 per million output tokens. Prompts exceeding 272,000 input tokens trigger a 2x rate on input tokens and a 1.5x rate on output tokens across the full request. Cache writes are billed at 1.25 times the uncached input rate.
In addition to direct API access, Sol is generally available on Amazon Bedrock running on its next-generation inference engine. AWS pricing matches first-party token rates, allowing usage to count toward existing cloud commitments. Regional processing endpoints with data residency controls and FedRAMP endpoints incur a 10 percent rate uplift.
Latest updates
Following general availability, a pricing update on August 21, 2026, reduced API and credit rates for Sol by over 20 percent for a three-month promotional period. An earlier update on July 30, 2026, renamed priority processing to Fast mode across the API ecosystem, with promotional rates active through November 30, 2026.
Execution speeds for Sol include Standard, Batch, Fast, and Ultrafast options. Fast mode input tokens cost $4.00 per million and output tokens cost $20.00 per million, whereas Ultrafast processing charges $12.00 per million input tokens and $60.00 per million output tokens. Batch processing provides reduced pricing for off-peak workloads.
Frequently asked questions
- When does OpenAI GPT‑5.6 Sol 2026 release?
- OpenAI GPT‑5.6 Sol 2026 is scheduled to release on Thursday, July 9, 2026 (exact time not yet confirmed).
- What is the context window capacity of GPT-5.6 Sol?
- GPT-5.6 Sol supports a context window of up to 1,050,000 input tokens and a maximum output limit of 128,000 tokens.
- Which input and output modalities are supported?
- The model accepts text and image inputs while generating text outputs. Audio and video modalities are not supported.
- Does GPT-5.6 Sol support fine-tuning?
- Fine-tuning is currently not supported for GPT-5.6 Sol.
- What reasoning effort levels can be specified?
- Reasoning effort levels can be set to none, low, medium, high, xhigh, or max, with medium acting as the default setting.
Last updated: 11/14/2026





