What changed
DeepSeek released V3.1 with a hybrid mode that supports both deliberate reasoning and direct responses, alongside improved tool use and updated API behavior.
Why it matters
A single model could serve both low-latency and reasoning-heavy work, reinforcing the shift from separate model families toward controllable inference modes.
What to do
Route classification and extraction through non-thinking mode, and reserve thinking mode for disputed claims, planning, or complex tool sequences; evaluate both modes at fixed cost and latency budgets.
Caveat
DeepSeek's performance and tool-use gains were first-party claims, and application teams still needed independent safety, citation, and tool-execution tests.