OpenAI launches GPT-6 Sol and Luna models with slashed API prices
OpenAI introduced GPT-6 Sol and Luna with 50% lower API costs than predecessors, offering specialized pricing for complex tasks and high-volume work.
- Headline: OpenAI launches GPT-6 Sol and Luna models with slashed API prices
- Dispatch Summary: OpenAI introduced GPT-6 Sol and Luna with 50% lower API costs than predecessors, offering specialized pricing for complex tasks and high-volume work.
- Verification: Corroborated across independent reporting outlets with primary sources and real-time wire transmissions.
OpenAI Unveils GPT-6 Sol and Luna: A 50% API Price Cut for Enterprise and Developers
On September 22, 2026, OpenAI introduced two new AI models—GPT-6 Sol and GPT-6 Luna—marking a significant shift in its pricing strategy and product lineup. The models, available via API and integrated into ChatGPT Work and Codex, offer 50% lower costs for developers and enterprises compared to their predecessors, GPT-5.6 Sol and Luna. This move positions OpenAI to compete more aggressively with rivals like Anthropic, which launched its Claude Opus 5.5 model the same day.
| Model | Input Token Price | Output Token Price | Price Change from GPT-5.6 |
|---|---|---|---|
| GPT-6 Sol | $2/million | $10/million | 50% reduction |
| GPT-6 Luna | $0.10/million | $0.50/million | 50% reduction |
Model Specialization: Sol for Complex Tasks, Luna for High-Volume Work
OpenAI positioned GPT-6 Sol as a "reasoning model" optimized for complex tasks like coding and professional workflows, while GPT-6 Luna targets high-volume, low-latency operations such as document summarization and information extraction. Both models share the same 1.05 million token context window and 128,000 token maximum output, but differ sharply in pricing and use cases.
"GPT-6 Sol makes about half as many mistakes as GPT-5.6 Sol," according to Macrumors, which noted improvements in coding accuracy and alignment with Astra's communication style. Luna, meanwhile, offers 90% discounts on cached input-token reads, enabling faster processing for repetitive tasks. However, OpenAI cautioned that Luna's lower cost comes with trade-offs: its knowledge cutoff is April 20, 2026, compared to May 18, 2026, for Sol.
Competitive Benchmarks and Rival Claims
OpenAI claimed GPT-6 Sol and Luna outperform Anthropic's Claude Opus 5.5 in benchmarks, but Anthropic disputed this, highlighting its newer Opus 5.5 model's superior performance on coding and knowledge work tasks. Mashable reported that GPT-6 Luna exceeded GPT-5.6's results by 5.4% on AutomationBench while costing 58% less per task. However, Anthropic's Opus 5.5 uses fewer tokens per task, complicating direct comparisons.
The pricing disparity between Sol and Luna has sparked debate among developers. Note.com calculated that processing 10,000 items monthly with 2,000 input and 500 output tokens would cost $90 for Sol versus $4.50 for Luna. This 20x difference underscores OpenAI's strategy to segment its user base: Luna for cost-sensitive applications, Sol for precision-critical workflows.
Availability and Regional Considerations
GPT-6 Sol and Luna are available to ChatGPT Work and Codex users on paid plans, with Luna extending to Free and Go tiers. However, neither model is accessible in basic Chat mode. The API rollout, which includes Japan in OpenAI's supported countries, remains unverified for regional-specific features. Note.com highlighted that while Japan is listed as a supported region, there is no explicit mention of Sol or Luna being targeted there.
Frequently Asked Questions
What is the price difference between GPT-6 Sol and Luna?
GPT-6 Sol costs $2 per million input tokens and $10 per million output tokens, while GPT-6 Luna is priced at $0.10 and $0.50, respectively. This represents a 20x cost difference for both input and output tokens.
Which models are available in ChatGPT and which require API access?
GPT-6 Luna is available to Free and Go users in the ChatGPT desktop app, while GPT-6 Sol is restricted to Work and Codex plans. Both models are accessible via API for developers, with no free tier available for API usage.
How do GPT-6 Sol and Luna compare to Anthropic's Claude Opus 5.5?
OpenAI claims GPT-6 Sol and Luna outperform Claude Opus 5.5 on benchmarks, but Anthropic asserts its Opus 5.5 model uses fewer tokens per task, making direct comparisons complex. Both companies highlight their models' strengths in coding and efficiency.
The launch of GPT-6 Sol and Luna signals OpenAI's focus on expanding access to advanced AI while maintaining cost-efficiency. As enterprises adopt these models, the next critical development will be how competitors like Anthropic respond to the pricing and performance dynamics. OpenAI's future announcements on model updates and regional expansions will likely shape the next phase of the AI market race.
How significant is this wire dispatch?
Cast your anonymous vote to register reader consensus across journalism and intelligence sectors.
Dateline Wire is dedicated to independent, evidence-backed reporting. This briefing was synthesized from primary source reporting, corroborated across independent newsrooms, and verified against our Editorial Standards.
