DeepSeek V4 Flash now runs updated weights on AI Gateway
By Steven Van ·
DeepSeek V4 Flash now runs on updated weights by default on AI Gateway, with notably stronger agentic capabilities.
DeepSeek V4 Flash now runs on updated weights by default in Vercel's AI Gateway, with stronger agentic performance. On Terminal-Bench, the model now scores 82.7, up 25.8 points from 56.9 in the April preview. Existing requests to deepseek/deepseek-v4-flash pick up the new weights automatically, with no change to the model ID or code.
DeepSeek is currently the only provider serving the updated weights; other providers, including ones with Zero Data Retention, are due next week. To run the model in a coding agent, use vercel ai-gateway coding-agents setup to connect the agent to AI Gateway and select deepseek/deepseek-v4-flash in its model configuration. As with other AI Gateway models, pricing matches the provider's rates with no markup and no platform fee, including on Bring Your Own Key requests.
