New Models
AI Minute Newsroom
2026-08-13
DeepSeek shipped its flagship with no announcement — same 1.6 trillion parameters, five times the agent score
Late on Wednesday night in Beijing, DeepSeek's API documentation quietly switched from the V4-Pro preview to a build named DeepSeek-V4-Pro-0813. There was no blog post and no launch event. It is the first finished version of the flagship that had been in preview since 24 April, and the architecture did not change: 1.6 trillion parameters in a mixture of experts, about 49 billion active per token, a million tokens of context, MIT licence. Everything moved in post-training, and it moved most where agents work. On DeepSeek's own scorecard DeepSWE goes from 12.8 to 62.7, Terminal Bench from 72.1 to 87.9, Cybergym from 52.7 to 83.3 and NL2Repo from 38.5 to 61.5. The API now also speaks the Responses and Anthropic protocols, so it drops into Codex and Claude Code without a wrapper. Pricing holds at ¥3 per million input tokens and ¥6 per million output — with a note from the company that it will rise significantly before long.
Why it mattersThe interesting part is not the size, because the size is unchanged. A model that scored 12.8 on an agentic coding benchmark in April scores 62.7 in August with the same parameter count and a different post-training run. That says most of the remaining headroom in agents is in training method, not in how many weights you can afford. For anyone paying to run a coding agent, the practical figure is the bill: roughly $0.84 per million output tokens, under a seventh of what the closed frontier charges, under an MIT licence that lets you host it yourself. DeepSeek has said in the same breath that the price will not stay there.
✓ Verified · 4 sources
Read in the app — free, in 9 languages
Related stories
OpenRouter dropped three AI providers after a bait-prompt leak test.
2026-11-22Researchers say they broke out of Meta's Muse agent and leaked chats.
2026-11-22Tencent released the weights of its flagship Hy4 model for download.
2026-11-21Amazon says attackers' AI agents found an unknown flaw on their own.
2026-11-21METR says it can rarely find humans who beat top AI at coding tasks.
2026-11-21