ITHome News from October 28: last night, a developer discovered a set of previously unseen DeepSeek-V5 model identifiers in the official DeepSeek Harness GitHub repository.
In the provider.spec.ts test file in the deepseek-harness repository, three entries — deepseek-v5, deepseek-v5-open, and deepseek-v5-next — appear near line 154, alongside the existing deepseek-v4-pro.
Judging from the code structure, these entries currently serve as framework-level placeholders rather than directly callable model interfaces. The v5-open entry carries requires ('0.0.1'), meaning this path has no substantive version gate, so legacy clients can continue to use it after the server-side switch; v5-next, meanwhile, is marked requires ('99.0.0-alpha.1'), pointing to a channel that requires a client upgrade before access. This shows that DeepSeek has already prepared two delivery schemes for V5 in its release mechanism: a "forced update" and a "smooth transition."
On April 24, 2026, DeepSeek released the V4 series in preview form, split into two versions, V4-Pro and V4-Flash, with V4-Pro entering officially usable status on August 13. In September, DeepSeek launched V4.1-Flash, which the company said uses an entirely new architecture, though the naming remains within the V4 series.
IT之家 reminder: the appearance of a placeholder does not mean a release is imminent. The existing code only proves that the "release mechanism is ready"; DeepSeek can enable V5 directly after modifying the / models configuration on the server side, without the client needing to release a new version.
As of October 28, DeepSeek has not officially issued any announcement, model card, or API identifier for the V5 series, nor has it made any public statement about the placeholder in the code repository.