DeepSeek V4-Pro Officially Launches with Major Agent Upgrades and Peak/Off-Peak Pricing
op7418 · x · 2026-08-13
DeepSeek's official API documentation reveals that the DeepSeek-V4-Pro official version is now live across app, web, and API, showing massive improvements in various Agent benchmarks like HLE and Terminal Bench. Meanwhile, the DeepSeek-V4-Flash official API has also entered public beta.
Key updates include:
- Enhanced Agent Capabilities: V4-Pro shows significant performance boosts in production environments and scores high on internal full-stack and hard coding benchmarks.
- Native Responses API Support: The API now natively supports the OpenAI Responses API format and is specifically optimized for Codex.
- Flexible Thinking Modes: Both V4-Pro and V4-Flash thinking modes now support three intensity levels (low, high, max) for flexible task management.
- Peak/Off-Peak Pricing: To optimize resource allocation, a new peak and off-peak pricing model will be introduced, with off-peak prices at 50% of peak rates, effective August 17, 2026.
More from coding & agent
- holaOS: Open-source all-in-one AI agent workspace — holaboss-ai · 2026-08-13
- Obsidian Skills: Empower AI agents to use Obsidian — kepano · 2026-08-13
- HolyClaude: One-Command Deployment for a Full AI Coding Workstation — tom_doerr · 2026-08-13
- Fine-tuning Inkling-Small on Twitter Data to Clone Personal Writing Style — IgorBrigadir · 2026-08-13
- City2Graph: Open-Sourcing Python Library for Urban Heterogeneous GNNs — Tough_Ad_6598 · 2026-08-13
- Anthropic Research: Identical AI Agents Converge on Same Bad Decisions, Causing Systemic Failures — rohanpaul_ai · 2026-08-13