DeepSeek V3.2 on serverless, plus cached and uncached token pricing in Model Library

☁️ Serverless

• DeepSeek V3.2 is now available on serverless

Cached Token Pricing Display

The Model Library and model detail pages now display cached and uncached input token pricing for serverless models that support prompt caching. This gives you better visibility into potential cost savings when using prompt caching with supported models.

Evaluations Dashboard Improvements

The Evaluations dashboard has been enhanced with new filtering and status tracking capabilities:

• Status column showing evaluator build state (Active, Building, Failed)
• Quick filters to filter evaluators and evaluation jobs by status
• Improved table layout with actions integrated into the status column

✨ New Models

• DeepSeek V3.2 is now available in the Model Library
• Ministral 3 14B Instruct 2512 is now available in the Model Library
• Ministral 3 8B Instruct 2512 is now available in the Model Library
• Ministral 3 3B Instruct 2512 is now available in the Model Library
• Mistral Large 3 675B Instruct is now available in the Model Library
• Qwen3-VL-32B-Instruct is now available in the Model Library
• Qwen3-VL-8B-Instruct is now available in the Model Library

📚 Documentation Updates

• Reranking Guide: Added documentation for using the /rerank endpoint and /embeddings endpoint with return_logits for reranking, including parallel batching examples (Querying Embeddings Models)

[accordion] Bug Fixes & Minor Improvements

• Deployment Page Enhancements: Redesigned deployment detail page with new model header, quick actions (Playground, Go to Model), tabbed API/Info interface, collapsible code examples, and improved GPU count display (Web App)
• Login Page Redesign: New marketing panel with customer testimonials carousel, highlighted platform capabilities, and improved visual design (Web App)
• Playground CTA: Added "Deploy on Demand" button next to "Try the API" for eligible models, making it easier to deploy models directly from the playground (Web App)
• Console Navigation Icons: Updated sidebar icons with a refreshed icon set for improved visual consistency (Web App)
• Reinforcement Fine-Tuning Defaults: Changed default epochs to 1 and increased maximum inference N from 8 to 32 for rollout configuration (Web App)
• Failed Job Visibility: Training progress and loss curves now display for failed and cancelled fine-tuning jobs, helping with debugging (Web App)
• Large Dataset Uploads: Improved upload handling for large JSONL files with progress tracking and direct-to-storage uploads (Web App)
• Dataset Preview: Fixed page freeze issue when previewing datasets with very long text content (Web App)
• Loss Chart Y-Axis: Fixed y-axis scaling on loss charts to properly display the full range of values (Web App)
• Model Deletion Dialog: Improved custom model delete confirmation dialog with better validation and feedback (Web App)
• Billing Date Range: Fixed date range calculation errors on the first day of the month in billing usage views (Web App)
• Login Session: Fixed an issue where expired sessions required manual cookie clearing to log in again (Web App)
• Rollout Detail Panel: Improved rollout log viewing with resizable split panels and better log formatting (Web App)
• Checkpoint Promotion: Added validation and error messages when promoting checkpoints with missing target modules or base model (Web App)
• Model Validation: Added validation before deploying fine-tuned models to ensure the model ID is valid (Web App)
• Quota Error Message: Improved error message clarity when request quota is exceeded in the playground (Web App)
• Safari Layout: Fixed extra spacing in login page marketing panel on Safari browsers (Web App)