DeepSeek-V5 placeholder appears in Harness code base, paving the way for a new generation of large models in advance

IT House reported on October 28 that yesterday evening, a developer discovered a set of DeepSeek-V5 model identifiers that had not appeared before in the official GitHub code base of DeepSeek Harness. In the provider.spec.ts test file in the DeepSeek-harness warehouse, 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 items currently play a placeholder role at the framework level, rather than directly callable model interfaces. The v5-open entry is accompanied by requires ('0.0.1'), which means that this path has no substantial version threshold, and old clients can continue to use it after the server is switched; v5-next is marked with requires ('99.0.0-alpha.1'), which points to a channel that requires the client to upgrade before it can access. This shows that DeepSeek has prepared two delivery solutions for V5, "forced update" and "smooth transition" in terms of release mechanism. On April 24, 2026, DeepSeek released the V4 series as a preview version, divided into two versions: V4-Pro and V4-Flash. V4-Pro entered the officially available state on August 13. In September, DeepSeek launched V4.1-Flash. Officially, it said it adopted a new architecture, but the naming was still classified as V4 sequence. IT House reminds: The appearance of placeholders does not mean that the release is approaching. The existing code can only prove that "the release mechanism is ready". DeepSeek can directly enable V5 after modifying the /models configuration on the server side. The client does not need to release a new version. As of October 28,