Skip to content

fix(provider): add NVIDIA NIM DeepSeek request compatibility #37833

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
fuselayer wants to merge 2 commits into
base: dev
Choose a base branch
from
+17 −0
Open

fix(provider): add NVIDIA NIM DeepSeek request compatibility #37833

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f7eacc4
fix(provider): remap developer role for NVIDIA NIM
fuselayer Jul 19, 2026
4399657
fix(provider): add NVIDIA NIM DeepSeek chat template kwargs
fuselayer Jul 19, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions packages/opencode/src/provider/transform.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,6 +79,13 @@ function normalizeMessages(
model: Provider.Model,
_options: Record<string, unknown>,
): ModelMessage[] {
// NVIDIA NIM rejects developer messages when chat_template_kwargs is present.
// Use system role instead for NVIDIA-hosted OpenAI-compatible models.
if (model.providerID === "nvidia") {
msgs = msgs.map((msg) =>
(msg as { role?: string }).role === "developer" ? ({ ...msg, role: "system" } as ModelMessage) : msg,
)
}
const sanitizeToolResultOutput = (content: ToolResultPart) => {
if (content.output.type === "text" || content.output.type === "error-text") {
content.output.value = sanitizeSurrogates(content.output.value)
Expand Down Expand Up @@ -1219,6 +1226,16 @@ export function options(input: {
result["enable_thinking"] = true
}

// NVIDIA NIM DeepSeek V4 requires chat_template_kwargs to stream reasoning.
if (input.model.providerID === "nvidia" && input.model.api.npm === "@ai-sdk/openai-compatible") {
if (modelId.includes("deepseek-v4")) {
result["chat_template_kwargs"] = {
thinking: true,
reasoning_effort: "high",
}
}

}
if (input.model.api.npm === "@ai-sdk/azure" && input.model.api.id.includes("gpt-5.5")) {
result["reasoningSummary"] = "auto"
return result
Expand Down
Loading