GPT-5.6 Sol Launches as OpenAI’s Most Powerful AI Yet, Challenging Claude Fable 5 at One-Third the Cost

OpenAI has unveiled GPT-5.6, its next-generation family of AI models led by GPT-5.6 Sol, a flagship model that delivers near-Claude Fable 5 intelligence while costing roughly one-third as much, according to independent benchmarks from Artificial Analysis.

The GPT-5.6 family introduces three new models—Sol, Terra, and Luna—designed to give developers and businesses more flexibility across intelligence, speed, and cost.

The announcement comes as competition between OpenAI, Anthropic, Google, xAI, and other frontier AI companies intensifies, with pricing and real-world performance becoming just as important as benchmark scores.

GPT-5.6 Sol Nearly Matches Claude Fable 5

Independent testing by Artificial Analysis places GPT-5.6 Sol just one point behind Claude Fable 5 on the Artificial Analysis Intelligence Index.

While Claude Fable 5 scored 60 points, GPT-5.6 Sol achieved 59 points—but at approximately one-third of the cost per task.

Artificial Analysis estimates GPT-5.6 Sol costs around $1.04 per benchmark task, compared to significantly higher costs for Claude Fable 5.

OpenAI also introduced:

• GPT-5.6 Terra for balanced everyday workloads
• GPT-5.6 Luna as the fastest and lowest-cost model in the family

Together, the three models offer different trade-offs between intelligence, reasoning speed, and operating costs.

New Leader in AI Coding

GPT-5.6 Sol also topped the Artificial Analysis Coding Agent Index, outperforming competing frontier models in coding evaluations including:

• DeepSWE
• Terminal-Bench v2
• SWE-Atlas-QnA

According to Artificial Analysis, GPT-5.6 Sol achieved an overall score of 80, leading every benchmark while remaining cheaper to run than both Claude Fable 5 and Claude Opus 4.8.

OpenAI says GPT-5.6 Sol establishes a new state of the art on Terminal-Bench 2.1, which measures long-running coding workflows involving planning, debugging, and tool coordination.

New “Ultra” Reasoning Mode

One of GPT-5.6’s biggest additions is a new Max Reasoning level alongside an Ultra mode.

Rather than relying on a single reasoning process, Ultra mode coordinates multiple AI sub-agents to tackle complex tasks requiring deeper planning and analysis.

OpenAI says the feature is designed for software engineering, cybersecurity, scientific research, and other long-horizon workflows.

Stronger Cybersecurity Capabilities

GPT-5.6 also improves significantly in cybersecurity evaluations.

According to OpenAI, Sol demonstrates stronger performance in vulnerability discovery, exploit analysis, and defensive security research while using fewer tokens than previous models.

The company emphasized that GPT-5.6 remains below its internal Cyber Critical threshold and cannot autonomously generate complete real-world attack chains under testing conditions.

Instead, OpenAI says the model is intended to help security professionals identify vulnerabilities, review code, and develop software patches.

Stronger Safety Before Public Release

Unlike previous launches, GPT-5.6 begins as a limited preview.

OpenAI says it worked closely with U.S. government officials before release and will initially make the models available only to a small group of trusted partners before expanding access in the coming weeks.

The company says it spent:

• More than 700,000 A100-equivalent GPU hours on automated red-teaming
• Extensive third-party security testing
• Multiple weeks hardening safeguards against jailbreak attempts

The models include stronger protections for cybersecurity, biology, and other high-risk domains.

Pricing

OpenAI also introduced a new pricing structure.

Model	Input Tokens	Output Tokens
GPT-5.6 Sol	$5 / million	$30 / million
GPT-5.6 Terra	$2.50 / million	$15 / million
GPT-5.6 Luna	$1 / million	$6 / million

The company also introduced cache-write pricing for the first time, while maintaining a 90% discount for cached reads.

Limited Preview Before Wider Release

During the preview period, GPT-5.6 Sol, Terra, and Luna will be available through the API and Codex for selected organizations.

OpenAI says broader availability across ChatGPT and the API is expected in the coming weeks after additional testing.

Frequently Asked Questions

1. What is GPT-5.6 Sol?

GPT-5.6 Sol is OpenAI’s flagship AI model in the new GPT-5.6 family. It offers advanced reasoning, coding, cybersecurity, and scientific capabilities while being more cost-efficient than previous OpenAI models.

2. What are GPT-5.6 Terra and GPT-5.6 Luna?

GPT-5.6 Terra is designed as a balanced model for everyday AI tasks, while GPT-5.6 Luna is OpenAI’s fastest and most affordable GPT-5.6 model, offering strong performance at a lower price.

3. How does GPT-5.6 Sol compare to Claude Fable 5?

According to Artificial Analysis, GPT-5.6 Sol scores just one point behind Claude Fable 5 on its Intelligence Index while costing approximately one-third as much per task, making it a highly competitive alternative.

4. Is GPT-5.6 Sol available to everyone?

No. GPT-5.6 Sol is currently in a limited preview for selected trusted partners through the OpenAI API and Codex. OpenAI plans to make it broadly available in ChatGPT and the API in the coming weeks.

5. How much does GPT-5.6 cost?

OpenAI prices GPT-5.6 Sol at $5 per million input tokens and $30 per million output tokens. GPT-5.6 Terra costs $2.50/$15, while GPT-5.6 Luna costs $1/$6 per million input and output tokens respectively.

6. What makes GPT-5.6 Sol better for coding?

GPT-5.6 Sol leads the Artificial Analysis Coding Agent Index and performs strongly on coding benchmarks such as Terminal-Bench v2, DeepSWE, and SWE-Atlas-QnA, making it one of the top AI models for software development.

7. What safety features does GPT-5.6 include?

GPT-5.6 introduces OpenAI’s strongest safety system to date, featuring layered safeguards, real-time cybersecurity and biology classifiers, automated red-teaming, jailbreak resistance, and enhanced protections against high-risk misuse.

Final Thoughts

GPT-5.6 represents one of OpenAI’s most significant model upgrades, focusing not only on stronger reasoning and coding performance but also on making frontier AI substantially more affordable. If independent benchmark results continue to hold in real-world deployments, GPT-5.6 Sol could become one of the strongest alternatives to Anthropic’s Claude Fable 5 for developers and enterprise customers looking to balance intelligence with cost.

Read More from Tech Category