Two models: Flash (284B total, 13B active) and Pro (1.6T total, 49B active). both hit 1M token context.

V4-Pro is their flagship. Beats Claude Opus 4.6 Max on Agent coding tasks (their words). specifically calls out being better than Sonnet 4.5 on coding, and competitive with Opus 4.6 on general benchmarks. on world knowledge and STEM, they say it's ahead of Gemini-Pro-3.1.

V4-Flash is the sleeper pick. Faster and cheaper than Pro, but it has better long-context efficiency than Pro does.

Original Text: Agent capabilities massively improved: V4-Pro hits SOTA on Agentic Coding benchmarks among open-source models. In practice, users report it feels better than Sonnet 4.5, and output quality is close to Opus 4.6 non-thinking mode — though there's still a gap vs Opus 4.6 with thinking enabled.

World knowledge: V4-Pro leads all open-source models by a significant margin on knowledge benchmarks, sitting just behind Gemini-Pro-3.1 among closed-source frontier models.

Top-tier reasoning: On math, STEM, and competitive coding, V4-Pro beats every open-source model that's been publicly benchmarked and is trading blows with the best closed-source models in the world. the 1M context is the real headline. Redesigned attention entirely — combines something called DSA (Deeply Sparse Attention) to handle the scale without blowing up compute. V4 inference cost stays flat as tokens scale up vs V3.2 which shoots up. the architecture improvement is what makes this actually usable, not just a spec number.

Agent capabilities got a dedicated upgrade. Trained specifically against Claude Code, OpenClaw, OpenCode, and CodeBuddy. V4-Pro is now the recommended model for any agentic / coding workflow. Flash is explicitly not recommended for the most complex agent tasks.

API is live. Pricing:

DeepSeek-V4-Flash: $0.14 / $0.28 per M input/output tokens

DeepSeek-V4-Pro: $1.74 / $3.48 per M input/output tokens

Reasoning_effort parameter lets you set thinking intensity (low/high/max) per call. "max" is recommended for agent tasks specifically.

The model will launch on Atlas Cloud. Developers can get API access.