Woofun AI reports that code identifiers for the DeepSeek V4 Flash vision model have been detected within the DeepSeek Harness repository. The specific tag 'deepseek-v4-flash-vision-exp' indicates that the latest Harness version is integrating support for native image requests.

However, the model has not been added to the official directory, and public API access attempts currently result in errors. Integration paths between the client and Harness are largely established, suggesting visual capabilities are nearing completion.