Off-peak rates cut 50%, flexible reasoning for Agents. OpenAI compatibility lowers switching friction further.
DeepSeek-V4-Pro GA Launch with Flexible Reasoning and Native OpenAI API Support
Original: DeepSeek-V4-Pro GA Release | DeepSeek API Docs
Importance: APIユーザーに即影響する価格改定と主要Agent機能追加、OpenAI互換性の実装により選定・運用に直結する
Summary
DeepSeek released V4-Pro GA on August 13, featuring major Agent upgrades and flexible reasoning modes (low/high/max) for both V4-Pro and V4-Flash. Native OpenAI Responses API support with one-click setup is now available. The model is accessible via "Expert Mode" on web/app and via API with unchanged model names. Concurrently, pricing shifts to peak/off-peak rates—off-peak 50% cheaper—effective August 16, 2026 at 16:00 UTC.
Key Points
- V4-Pro GA with enhanced Agent capabilities and reasoning flexibility
- Reasoning levels selectable (low/high/max) for task-specific optimization
- Native OpenAI Responses API with one-click Codex setup
- Dual pricing tiers from Aug 16: off-peak rates 50% cheaper than peak
View developer notes (APIs, breaking changes, migration)
V4-Pro API deployment preserves existing model names—reference API docs for setup details. Native OpenAI Responses API support with one-click Codex optimization. Reasoning complexity controlled via request parameters (low/high/max). Dual pricing tiers effective Aug 16, 2026, 16:00 UTC: peak vs. off-peak at 50% discount. Existing integrations remain functional; workload scheduling to off-peak windows is recommended for cost optimization.
Source: https://api-docs.deepseek.com/news/news260813
Outlet: DeepSeek Release Notes
This article is an AI-generated summary (OpenAI GPT-4o-mini) of publicly available information from Anthropic, OpenAI, Google, Meta, Mistral, DeepSeek, Sakana, and other vendors. The original source URL is always provided in accordance with fair-use citation requirements. Summaries are AI-generated and may contain mistranslations or misinterpretations. Always verify details with the original source.