Modèles OpenAI

For complex coding, computer use, and professional document work. It improves on GPT-6 Sol in demanding tasks and approaches Astra at a lower cost; choose Astra when maximum quality is the priority.

Built for complex coding and agentic workflows, including project-level programming, long-horizon planning, and multi-step tool use. Its 1.05M-token context, text/image input, reasoning, function calling, and structured outputs make it the workhorse choice; it costs more than GPT-6 Luna when quality and execution stability matter most.

Designed for focused, high-volume work such as batch summarization, classification, rewriting, and frequent agent calls. It still offers a 1.05M-token context, text/image input, reasoning, function calling, and structured outputs, but trades depth for much lower cost; prefer GPT-6 Sol for the hardest long-horizon tasks.

The discounted-price route for GPT Image 2.5 Sunburst, mapped by the current channel to Sunburst. Sale indicates a special-price route and does not change Sunbursts role in high-precision image generation and editing. Best for cost-sensitive professional visual work that still needs reference-image editing and detail control; use regular gpt-image-2.5-sunburst when you want the standard route.

Built for premium visual workflows where editing precision and finished-image quality matter most: a better fit for production campaign creative, polished product imagery, and multi-turn edits that need to preserve the subject, composition, and visual treatment. It accepts text and reference-image inputs and produces images, with longer generation times than Flare.

The discounted-price route for GPT Image 2.5 Flare, mapped by the current channel to Flare. Sale indicates a special-price route and does not change Flares role in fast, high-quality everyday image generation and editing. Best for cost-sensitive, frequent image tasks; use regular gpt-image-2.5-flare when you want the standard route.

A strong default for high-quality image generation and editing: optimized for faster iteration across creator and social content, product experiences, visual search, rapid prototyping, and high-volume generation. It accepts text and reference-image inputs and produces images; choose Sunburst when tighter editing control matters more than speed.

A flagship end-to-end model for complex reasoning, software engineering, computer use, deep research, and document work. It offers about a 1.05M-token context window, accepts text, images, and documents, returns text, and supports reasoning, search, function calling, structured outputs, code execution, and tool orchestration. Choose it over lower-tier models for long-horizon task quality and execution stability, with higher usage cost as the trade-off.

The balanced workhorse tier of GPT-5.6, combining capability and cost. Well suited to production work involving image understanding, tool use, and large source material; choose Sol for the hardest quality-first work or Luna for cost-controlled high volume.

The flagship GPT-5.6 tier for complex professional work and quality-first delivery. It supports image input, function calling, structured outputs, and an exceptionally large context window; consider Terra or Luna when latency or unit cost is the primary constraint.

The cost-sensitive, high-volume tier of GPT-5.6. A good fit for summarization, classification, rewriting, and batch automation; choose the Sol sibling for complex professional judgment or quality-first multi-step work.

Conçu pour le raisonnement, le codage, la compréhension visuelle et les tâches d'agent, avec un contexte long, une utilisation des outils et une sortie fiable et évolutive plus forts que GPT-5.4. Comparé aux modèles phares Claude, Gemini et Qwen, il s'agit d'une base polyvalente solide pour une automatisation fiable et des flux de connaissances complexes.

The discounted-price route for GPT Image 2, mapped by the current channel to the same underlying model. Sale indicates a special-price route, not a separate capability tier. A good fit for cost-sensitive everyday generation and reference-image editing; choose regular gpt-image-2 when route consistency matters more.

Conçu pour la génération et l'édition d'images de haute qualité, améliorant par rapport à GPT Image 1.x le rendu du texte, l'édition locale et la cohérence multi-tours. Comparé aux modèles d'image Gemini, Qwen Image et Seedream, il est bien adapté aux visuels de marque, à la conception publicitaire, à l'imagerie de produits et aux flux de travail créatifs étroitement contrôlés.

Built for reasoning, coding, visual understanding, and agent tasks, with stronger long context, tool use, and scalable reliable output than GPT-5.3. Compared with Claude, Gemini, and Qwen flagship models, it is a strong general-purpose base for reliable automation and complex knowledge workflows.

Conçu pour le raisonnement, le codage, la compréhension visuelle et les tâches d'agent, avec un contexte long, une utilisation des outils et une sortie fiable et évolutive plus forts que GPT-5.3. Comparé aux modèles phares Claude, Gemini et Qwen, il s'agit d'une base polyvalente solide pour une automatisation fiable et des flux de connaissances complexes.

Conçu pour le raisonnement, le codage, la compréhension visuelle et les tâches d'agent, avec un contexte long, une utilisation des outils et une sortie fiable et évolutive plus forts que GPT-5.3. Comparé aux modèles phares Claude, Gemini et Qwen, il s'agit d'une base polyvalente solide pour une automatisation fiable et des flux de connaissances complexes.