Vultr Serverless Inference: Outdated model names and missing model management UI

[DanielQura]

Description

The application lists outdated Vultr models that do not align with the current official Vultr Serverless Inference documentation. Additionally, the UI lacks functionality to manually add, remove, or edit model names to override the hardcoded list.

Identified discrepancies between UI and Vultr documentation:

UI: MiniMax M2.5 | Docs: MiniMax-M2.7

UI: Kimi K2 Instruct | Docs: Kimi-K2.6

UI: DeepSeek V3.2 | Docs: DeepSeek-V3.2-NVFP4

UI: GLM 5 FP8 | Docs: GLM-5.1-FP8

Missing from UI entirely: Llama-3.1-Nemotron-Safety-Guard-8B-v3, Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16, Nemotron-Cascade-2-30B-A3B

Plugins

cli and desktop app

OpenCode version

OpenCode Desktop v1.15.3

Steps to reproduce

1. Navigate to the Vultr provider model configuration settings.
2. Review the list of available models presented as toggle switches.
3. Compare the available models against the official Vultr Serverless Inference supported chat completion models.
4. Search for a UI element to add custom models or remove obsolete ones. Observe its absence.

Screenshot and/or share link

Operating System

Windows 11 Home - 25H2 - 26200.8457

Terminal

Terminal and Powershell