OpenAI rolled out its latest AI family on Thursday, introducing three new models—Sol, Terra and Luna—collectively branded as GPT-5.6. The company framed the launch as a leap in efficiency, highlighting token‑saving gains that translate into lower costs and faster results for users ranging from developers to security teams.
Sol, described as the workhorse of the set, delivers the highest performance. OpenAI claims it is 54% more token‑efficient on coding tasks than its predecessor, a figure the CEO, Sam Altman, repeated in a recent interview with CNBC. Terra offers an intermediate balance of speed and cost, while Luna targets budget‑conscious customers, delivering comparable capabilities at a fraction of the price.
Cybersecurity focus
The new family is marketed as OpenAI’s "strongest cybersecurity model yet." According to the company, the models excel at threat modeling, code review, patch generation and blue‑team simulations—essentially letting organizations probe their own defenses before real attackers strike. This angle comes amid lingering concerns from the Trump administration about potential misuse of advanced AI, which had previously sought to limit the rollout of powerful models.
Enterprise tools and pricing
Alongside the models, OpenAI introduced ChatGPT Work, a workplace companion built to run on desktop, web and mobile platforms. The tool aims to automate routine tasks such as drafting documents, building spreadsheets and creating presentations, positioning itself as a productivity booster for corporate teams.
Pricing reflects the tiered model approach. Sol costs $5 per million input tokens and $30 per million output tokens. Terra is priced at $2.50 for input and $15 for output, while Luna starts at $1 per million input tokens and $6 per million output tokens. The company says the models are available through ChatGPT, Codex and the OpenAI API.
OpenAI’s rollout arrives as competitors like SpaceXAI and Meta have also announced new offerings this week. However, the firm frames GPT-5.6 as a direct response to Anthropic, which has been gaining traction with enterprise customers. Using the Artificial Analysis Coding Agent Index, OpenAI asserts that Sol outperforms Anthropic’s Fable 5 by 2.8 points, using less than half the output tokens and costing about a third as much. Terra and Luna also claim modest advantages over Anthropic’s lineup.
Industry observers note that the emphasis on token efficiency and cost could make GPT-5.6 attractive to businesses seeking to scale AI applications without exploding budgets. The added focus on cybersecurity aligns with growing demand for AI‑driven defense tools, a market that has seen heightened interest following high‑profile cyber incidents.
OpenAI’s announcement underscores its strategy of bundling advanced model capabilities with practical enterprise tools, hoping to lock in larger corporate clients while staying ahead of rivals in both performance metrics and price points.
Este artículo fue escrito con la asistencia de IA.
News Factory APP - noticias agénticas para impulsar tu SEO y AEO.