1. State-of-the-Art Image Generation

• Produces high-quality, detailed images optimized for realism, style control and prompt fidelity.
• Designed to handle complex visual scenes, compositions and lighting conditions.

2. Natively Multimodal Architecture

• Understands and reasons over both text and images as inputs.
• Ideal for workflows like editing based on reference images, expanding sketches or mockups and visual concept development.

3. Flexible Output Resolutions & Quality Levels

• Supports multiple resolutions including 1024x1024, 1024x1536 and 1536x1024.
• Offers three quality tiers (Low, Medium, High) to balance cost, speed and maximum detail.

4. Multiple Pricing Models

• Pay-per-token for multimodal input: text tokens and image tokens.
• Pay-per-image generation for final output: low, medium and high quality tiers.
• Enables businesses to balance cost and output needs.

5. Broad Use Cases

• Product photography and marketing assets.
• Illustration, concept art and creative ideation.
• UX/UI mockups.
• Style-guided image creation.
• Generating reference images for design or storytelling.

6. Supported Across Major API Endpoints

• Available via Chat Completions, Responses, Realtime, Assistants and Images (generations/edits) endpoints.
• Allows tight integration into automated creative pipelines or user-facing apps.

7. Simplified Model Behavior for Stability

• No streaming, function calling, structured outputs or fine-tuning; focused solely on high-quality image generation.

8. Consistent Results via Snapshots

• Supports snapshots for version locking to ensure long-term reproducibility.

9. Ideal For

• Designers, marketers and creatives.
• Product teams needing image assets.
• App builders integrating image generation workflows.
• Agencies producing visual content at scale.

1. Intelligence & Benchmark Performance

• Matches Claude 3 Opus (previous largest model) on many intelligence tasks.
• Surpasses Claude 3 Opus on multiple evaluations despite being a smaller, faster model.
• Major improvements across every skill category vs previous Haiku.

2. Coding Strength

• Scores 40.6% on SWE-bench Verified, outperforming:

  • Claude 3.5 Sonnet (original version)
  • GPT-4o
  • Many agent-driven systems

• Excellent for engineering assistants, agent coding tasks, and bug fixing.

3. Speed & Latency

• Same speed class as Claude 3 Haiku (ultra-fast).
• Ideal for real-time interactions, high request volumes, and UI responsiveness.

4. Tool Use & Instruction Following

• Better at following instructions than previous Haiku.
• Stronger at tool use accuracy, making it reliable for agents and workflows.

5. Best Use Cases

• High-volume, low-latency tasks
• User-facing products
• Sub-agent tasks in larger workflows
• Processing large structured datasets (pricing, inventory, purchase history)
• Rapid content or code generation where speed matters