The latest OpenAI GPT models at a glance
(Last updated: August 2026)
What makes GPT-5 stand out?
OpenAI’s GPT-5 ushered in a new generation of powerful large language models Since its launch, a wide range of model variants have been released, offering different functions and performance modes (fast, strong reasoning, high-performance). GPT-5 is both a model name and the designation for a family of models, with the latest models in the family currently being GPT-5.6, comprising Sol, Luna and Terra. The new Frontier models promise improved token efficiency and enhanced cybersecurity and coding performance.
OpenAI’s aim is to make advanced intelligence more widely available and affordable. According to OpenAI, GPT-5.6 Luna in particular offers cutting-edge intelligence at a fraction of the cost of comparably powerful models. The model promises shorter response times and more stable performance, whilst also enabling longer dialogue sessions and personalised responses. With these expanded model variants, the GPT-5 series can be used in a wide variety of areas, ranging from content creation and programming to complex analytical tasks and cost-effective real-time applications.
Here is an overview of the key new features, based on official information from OpenAI:
- The current generation of GPT is based on the GPT-5 model family. OpenAI is developing various variants within this family; ChatGPT uses automatic model selection and system optimisations to provide suitable models and features depending on the task.
- The availability of individual models in ChatGPT changes regularly: OpenAI updates the supported models and features depending on performance and user requirements.
- In addition to new model generations, OpenAI is increasingly expanding ChatGPT with product features such as research, image processing, coding support and AI agents.
Versions and use of the GPT models
(Last updated: August 2026)
The latest flagship is the GPT-5.6 family (Sol, Terra, Luna), which was fully released on 9 July 2026. Sol is the most powerful model for complex workflows, coding, science and cyber security; Terra is a well-balanced model for everyday use; and Luna is the cost-effective option. GPT-5.5 Instant remains the standard model for everyday queries in ChatGPT, complemented by GPT-5.5 Thinking and GPT-5.5 Pro for more in-depth reasoning. Parts of the GPT-5.4 range (Thinking, Pro, mini, nano) as well as GPT-5.3 remain in use within the routing system.
GPT-5 represents the ongoing evolution of OpenAI’s AI models. GPT-4 laid the foundations with advanced text and image processing and reliable accuracy in everyday tasks. GPT-4o, as a turbo-optimised variant, operated faster and more accurately, and also experimented with audio and video processing, as well as empathetic and emotional tone. As the next generation, GPT-5 combines the highest accuracy with native multimodality for text, images, audio and video, an extremely large context window, and an integrated thinking mode. The choice of model depends on the required accuracy and capacity, depending on the use case.
Discontinued Models
Several older models have since been removed from ChatGPT: GPT-4o and the rest of the GPT-4 series (including GPT-4.1, GPT-4.1 mini, o4-mini and GPT-5 Instant/Thinking) as early as February 2026, GPT-5.1 (Instant, Thinking, Pro) in March 2026, GPT-5.2 in June 2026 and, most recently, GPT-4.5 at the end of June 2026. OpenAI is thus clearly focusing on the GPT-5.x generation and a more system-based, automatically routing model architecture. Since January 2026, ChatGPT Go has offered an affordable subscription tier (around $8 per month) between the Free and Plus tiers, which has significantly broadened access.
- GPT-5.3 / GPT-5.3 Codex (early 2026): Predecessor to 5.5 Instant, a specialised Codex variant for coding agents; snapshots will be phased out gradually from August 2026.
- GPT-5.2 (Dec. 2025): Improvements in structured reasoning; removed from ChatGPT in June 2026.
- GPT-5.1: Transitional model with adjusted response behaviour; removed in March 2026.
- GPT-5 (first generation): Introduced the adaptive Instant/Thinking model selection; removed from ChatGPT in February 2026.
GPT-5.6 vs. GPT-5.5
Whilst GPT-5.5 already represented a major leap forward in terms of accuracy, reliability and suitability for everyday use, GPT-5.6 is now setting new standards in precisely these areas of performance. As the standard model in ChatGPT, GPT-5.5 continues to offer developers and businesses a powerful platform for text processing, chatbots, content generation and everyday knowledge work. GPT-5.6 significantly expands on these capabilities: Sol, in particular, offers significantly improved capabilities in coding, scientific analysis and cyber security compared to GPT-5.5 and, according to OpenAI, is considered the most powerful model to date for demanding, multi-stage tasks. Furthermore, GPT-5.6 Sol introduces new control mechanisms such as a customisable ‘Max Reasoning Effort’ level and an ‘Ultra’ mode with multiple sub-agents working in parallel, making complex programming and analysis tasks noticeably more efficient than with the previous model.
Current debates surrounding GPT-5
Current debates surrounding OpenAI’s models now focus less on personality or emotionality and more, above all, on transparency and trust in relation to the GPT-5.6 release.
Security concerns
GPT-5.6 was initially made available only to a small group of partner organisations, following intervention by the US government due to the model’s significantly enhanced cybersecurity capabilities (particularly those of Sol). Furthermore, a hack carried out autonomously by the model has raised questions amongst critics as to whether OpenAI’s own security criteria are being consistently applied. OpenAI recently introduced ‘Advanced Account Security’. These controls are designed to strengthen login protection, tighten account recovery procedures, reduce the risk posed by compromised sessions, and provide users with greater transparency regarding account activity.
Confusion over product structure
OpenAI has also restructured its app landscape: the previous ChatGPT app is now called ‘ChatGPT Classic’, whilst the former Codex app is becoming the central desktop app for Chat, Work and Codex. This is causing uncertainty on forums such as Reddit, whilst the three-variant structure (Sol/Terra/Luna) is perceived as unnecessarily complicated. Terra’s value for money, in particular, is attracting a great deal of attention.
Advertising and user trust
As recently as 2024, Sam Altman described advertising as a “last resort” for ChatGPT. An official advertising trial has been running in the US since February 2026 and has been gradually expanded since then. Adverts are clearly labelled as “Sponsored” and visually separated below the ChatGPT response, displayed as “Sponsored Recommendations” relevant to the current conversation. However, this applies exclusively to Free and Go users in the US. Advertising is still not active in Germany or the DACH region, although the German ChatGPT Go pricing page already displays the note “This plan may contain advertising”.
Rapid model development and intense competition
The rapid succession of releases, from GPT-5 right up to version 5.6, within the space of a few months is causing increasing confusion amongst users and developers. The discussions that have emerged have centred primarily on the following questions:
- Are smaller, specialised variants more useful?
- Are new generations of models being released and replaced too soon?
- What about the quality and ethical considerations of the new models?
OpenAI’s market leadership is under pressure: a lack of transparency, distortion of competition and changes such as the introduction of advertising are leading users to increasingly turn to alternatives on the market, such as Google Gemini or Anthropic’s Claude. Anthropic is deliberately positioning its chatbot, Claude, as an ad-free alternative.
Assessment and Outlook
Despite impressive technical advances, the GPT-5 cycle demonstrates that OpenAI is increasingly facing growing competitive pressure from major rivals such as Google and declining trust amongst its own user base. Nevertheless, GPT-5 has now established itself as an integral part of numerous enterprise and developer platforms, including Microsoft Copilot and Azure AI Foundry. This widespread integration optimises the use of GPT-5 in business processes and customer service applications. The introduction of a hybrid multi-model system enables more efficient task allocation and leads to improved performance when handling complex queries. Benchmark tests confirm higher accuracy and reduced error rates compared with previous models.
Future models could feature improved personalisation capabilities to cater for individual user preferences, as well as further expand the context window and offer even more precise multimodality. Specific features have not yet been announced; we must await the official announcement of new models.
GPT-5 hat vor allem seine Stärken in den agentischen Anwendungen, besonders im Bereich der tieferen Recherche und als Assistent beim Programmieren.
What about gpt-oss?
Although it hasn’t generated much buzz in the wider news world, here at moinAI we’re watching the open-source model gpt-oss with great interest. gpt-oss-120b and gpt-oss-20b remain OpenAI’s only freely available models. The advantage of the OSS version for moinAI, as a SaaS provider, is that it is very fast and high-performing. This enables us to offer a high-quality service, hosted entirely on German servers at moinAI, without relying on providers such as OpenAI or Azure. This ensures that customer data is never passed on to any third parties.
Conclusion
GPT-5 already represents a significant step forward in AI development, but OpenAI must address the critical feedback and incorporate it into future releases. The biggest challenge for OpenAI is therefore to strike a balance between technical excellence and engaging user interaction.
[[CTA headline="Start your company's AI journey with moinAI!" subline="Unlock the full potential of modern language models and AI technologies in your business." placeholder="Insert your company URL here" button="Try it now!"]]
