OpenAI launches GPT-5.6 with better performance and reduced costs
OpenAI launches GPT-5.6 with better performance and reduced costs
OpenAI has announced the general availability of the GPT-5.6 family, with three models - Sol, Terra, and Luna - focusing on improved performance and lower costs. According to the company, its new flagship model, GPT-5.6 Sol, achieves top results in programming, knowledge work, cybersecurity, and science, using fewer tokens and at an estimated lower cost compared to previous and competitor models.
The launch follows a limited preview phase. OpenAI presents Sol as its most advanced model, Terra as a balanced version for daily work, and Luna as its most economical model. The company also introduces an ultra setting, described as its highest capacity level, which coordinates multiple agents in parallel to process complex tasks faster.
Gains highlighted on costs and speed
OpenAI claims to have trained GPT-5.6 to achieve more useful work per token. On Agents’ Last Exam, an evaluation of long-term professional workflows in 55 domains, GPT-5.6 Sol reaches 53.6. The company states that this surpasses Claude Fable 5 (adaptive reasoning) by 13.1 points. It adds that in average reasoning, the model stays ahead of Fable 5 by 11.4 points for about a quarter of the estimated cost.
According to OpenAI, this efficiency logic also extends to smaller models. GPT-5.6 Terra and GPT-5.6 Luna outperform Fable 5 at about one-sixteenth of the cost. On the Artificial Analysis Intelligence Index, GPT-5.6 Sol with maximum reasoning is less than a point away from Fable 5, while completing tasks 61% faster and at around half the estimated cost.
Regarding API pricing, OpenAI states that GPT-5.6 is billed per 1M tokens across three model sizes: Sol at $5 input / $30 output, Terra at $2.50 input / $15 output, and Luna at $1 input / $6 output.
sbb-itb-454261f
Programming, agents, and parallel execution
OpenAI claims that GPT-5.6 Sol is its best programming model to date. On the Artificial Analysis Coding Agent Index, GPT-5.6 Sol with maximum reasoning reaches 80, which is 2.8 points above Fable 5, while using less than half of the output tokens, taking less than half the time, and costing about a third less. Terra slightly outperforms Fable 5, and Luna surpasses Opus 4.8, according to the company's cited figures.
The company also highlights Programmatic Tool Calling in the Responses API. This feature allows the model to write and execute lightweight programs in memory to coordinate tools, process intermediate results, and adapt workflow on the fly.
For heavier tasks, OpenAI offers the max and ultra levels. According to the company, ultra defaults to coordinating four agents in parallel. In comparisons presented on BrowseComp, SEC-Bench Pro, and Terminal-Bench 2.1, OpenAI claims that adding parallel agents improves the score-latency ratio.
Knowledge work, design, and computer usage
OpenAI states that GPT-5.6 brings a leap in design judgment and in the production of interfaces, presentations, documents, and spreadsheets. The company says the model can transform natural language requests into interactive explanations and visualizations in ChatGPT Work, and produce fully editable presentations from source materials.
On BrowseComp, GPT-5.6 Sol achieves 92.2% with Ultra and 90.4% in the standard version, according to the provided figures. On OSWorld 2.0, Sol reaches 62.6% and surpasses Opus 4.8 using 85% fewer output tokens, as per OpenAI. The company adds that Luna nearly approaches the best performances of GPT-5.5 at less than half the estimated cost, while Terra exceeds them at a lower cost.
Results also highlighted in cybersecurity and science
OpenAI presents GPT-5.6 as its most performing model to date in cybersecurity. On ExploitBench, it achieves 73.5% compared to 47.9% for GPT-5.5 at a comparable budget in output tokens. On ExploitGym, the maximum success rate increases from 15.1% with GPT-5.5 to 24.9% within two hours, then to 33.7% within six hours. On SEC-Bench Pro, GPT-5.6 Sol reaches 71.2% compared to 45.8% for GPT-5.5.
In scientific research, OpenAI claims that GPT-5.6 Sol shows extensive gains. On GeneBench Pro, it achieves 28.7% compared to 12% for GPT-5.5. The company also mentions improvements in biology, life sciences, and chemistry.
Security and deployment
OpenAI claims to have deployed its strongest safeguards to date for GPT-5.6. The company says it combined human red teaming, large-scale automated testing, and around 700,000 A100e GPU hours of automated black-box red teaming before general availability.
According to OpenAI, GPT-5.6 is more capable than previous models in biology and cybersecurity without crossing the Critical threshold in any of these domains. The company also states that the cybersecurity safeguards of GPT-5.6 Sol block about ten times more potentially harmful activities than those of previous models.
Availability
GPT-5.6 is available starting today in ChatGPT, Codex, and the OpenAI API. OpenAI specifies that the deployment is starting globally and is expected to gradually progress towards full availability over the next 24 hours.
What the first users and partners say
"GPT-5.6 is one of the strongest models we’ve tested on CursorBench, delivering solid results in early evals. It’s an exciting step forward for developers for persistence, intelligence, and overall efficiency. We are looking forward to bringing this model to our Cursor users."
- Oskar Schulz, President at Cursor
"GPT-5.6 was the strongest model we evaluated on our agentic code-review tests. On our apples-to-apples internal and external PR benchmarks, it beat GPT-5.5 on F1 while using roughly 3x fewer tokens per PR and delivering about 2x lower median latency."
- Itamar Friedman, Co-Founder & CEO at Qodo
"GPT-5.6 Sol is really, really good. It’s the most tenacious problem-solver we’ve seen yet, staying focused and on-task for days at a time. It’s exceptional at updating Custom Agents and refining memories as your workspace evolves, so they get sharper the longer they run. Terra and Luna also punch well above their price. Many agents running GPT-5.5 perform just as well on Terra for half the cost and 16% fewer tokens."
- Simon Last, Co-Founder at Notion
"For production coding agents, GPT-5.6 stood out as a top-tier model that combines strong coding-agent performance with very strong cost efficiency."
- Scott Wu, Co-founder & CEO at Cognition
"GPT-5.6 is a major step forward for financial research agents. On Rogo’s Big Finance Benchmark, it improved rubric quality by 6.2 points and answer accuracy by 3.6 points versus GPT-5.5. With Programmatic Tool Calling, it matched quality while using 24% fewer output tokens and completing tasks 28% faster. That combination of accuracy, speed, and efficiency is exactly what we need to scale high-quality financial analysis."
- Alex Wang, Applied AI at Rogo
"GPT-5.6 felt less like a chat assistant and more like an end-to-end technical operator. It could inspect live systems, debug issues, make code changes, validate results, publish artifacts, and carry context across long sessions with strong grounding."
- Ian Tracey, Software Engineer, Applied AI at Ramp
"GPT-5.6 was much better than predecessors at understanding the layer of work I wanted. Across a multi-stage Codex workflow of research, planning, then staged implementation, it followed intent better than GPT-5.5, and consistently produced accurate line-linked GitHub references where prior models often missed."
- Shane Moran, Senior Applied AI/ML Engineer at Shopify
"GPT-5.6 consistently stays focused through long-running tasks, makes excellent use of tools, and gets to high-quality solutions with little steering. For research and design work, it produces clear reports and intuitive diagrams that help our teams understand complex systems and move faster."
"Across legal research and document workflows, GPT-5.6 is already delivering the kind of efficiency gains that change product economics. In our combined evaluation suite, it uses 14% fewer tokens while improving quality across legal research and transactional law use cases. For multi-step document analysis, Programmatic Tool Calling cuts prompt tokens by 38% with no quality loss."
- Angel Faus, VP of Engineering at Clio
"GPT-5.6 delivered the best efficiency profile we’ve seen for complex financial research. In our evals, it performed at a top-tier level while being 1.72x more token-efficient, leading in three headline categories and scoring 88% on multi-hop tasks. The combination of efficiency, accuracy, and quality makes the model a good fit for scaling financial research workflows."
- Alberto Da Costa, Principal Engineer, Applied AI at Balyasny Asset Management
"GPT-5.6 Sol showed substantial improvements on reasoning, decision making, and autonomy. The improvements to subagent use are particularly valuable for complex accounting work. Excited for the direction of agent development for OpenAI."
- Tarrek Shaban, Head of Product at Basis
"GPT-5.6 is notably efficient on the long, complex workflows behind building production-grade apps. As one of the models now used by Lovable, it delivers for users with roughly 25% fewer steps and 35–48% fewer tool calls than the prior model, while improving project success and reducing stuck runs by 15%. That’s a meaningful difference for anyone trying to go from idea to working app."
- Fabian Hedin, Co-Founder at Lovable
"GPT-5.6 Sol is the first model we’ve evaluated that consistently generates decks ready for real work. Across 20 challenging client workflows and hundreds of decks in Model ML’s FinBench, it used 39% fewer tokens per deck than Fable while producing more polished, legible decks with clearer, more accurate data visualizations that required less rework before sharing."
- Chaz Englander, Co-Founder & CEO at Model ML
"GPT-5.6 was the best overall frontend model in our seven-task benchmark. On our five-point frontend QA rubric, it scored 4.4, compared with 4.0 for GPT-5.5 and 3.5 for Claude 4.8, and consistently turned complex ecommerce, dashboard, and product briefs into complete, responsive interfaces across desktop and mobile."
- AJ Orbach, CEO at Triple Whale
"With GPT-5.6’s Programmatic Tool Calling, we could build detailed Unity scenes through our structured API much more efficiently. Across scene-construction workflows, it used 63.5% fewer total tokens and 50.1% fewer model turns than the same model using direct tool calls, while producing comparable visual results. That makes iterative game creation much more practical to scale."
- Teddy Cross, Co-Founder & CPO at PlayCo
"GPT-5.6 is especially strong on presentations. In our early design evals, it was stronger than competitive models for slide creation and about 1.6x more token-efficient, which matters when you’re generating and refining visual work at Canva scale."
- Danny Wu, Head of AI Products at Canva
"GPT-5.6 marks an advancement for artifact generation in Microsoft 365. In our evaluations, it delivered strong results across a wide range of productivity scenarios, producing outputs that were highly cohesive, accurate, and ready for use. By reducing the effort required to refine prompts and iterate on drafts, it helps users spend less time shaping content and more time acting on it."
- Charles Lamanna, EVP, Copilot, Agents and Platform at Microsoft
"Across 30 real-world app-building conversations, GPT-5.6 used 22% fewer input tokens and 23% fewer output tokens than GPT-5.5 while staying competitive on greenfield and long multi-turn work. It’s a genuine step up, especially in design and frontend capabilities."
- Gabriel Grinberg, AI Engineering Lead at Base44
"GPT-5.6 was a step up on legal workflows. In Legora’s internal eval harness, it improved or held steady in 5 of 7 tasks, with the strongest gains in structured drafting and precedent review, while staying appropriately cautious on legal conclusions."
- Jake Lauritzen, CTO at Legora
"With GPT-5.6 in Figma Make, teams can turn even complex designs into interactive prototypes. It raises the bar for design-to-code workflows."
- Loredana Crisan, Chief Design Officer at Figma
OpenAI also states that GPT-5.6 accelerates its own research work. During the internal testing period, the daily average of output tokens per active researcher was more than twice the highest level observed with GPT-5.5. The company further indicates that over the past six months, the share of research computation dedicated to internal code inference has increased by a factor of 100, while the internal use of agent tokens has increased by about 22 times.
The launch of GPT-5.6 thus marks, according to OpenAI, a combination of higher performance, reduced costs, and strengthened safeguards across its Sol, Terra, and Luna models.