Add NVIDIA NIM platform (provider 11)

The 14th registry row: id "nvidia", NVIDIA_API_KEY > auth.json "nvidia"
scope, https://integrate.api.nvidia.com/v1 with KIGI_NVIDIA_BASE_URL
override, Bearer, ChatCompletions, enrichment-backed metadata
(models_dev_id nvidia — the quirk matrix's earlier 'absent' claim was
wrong; models.dev has 84 nvidia models), restrict_to_enriched=true (the
NIM listing mixes chat/embedding/rerank/vision/image; keep the 45
tool-calling chat models). Slashed org/model ids
(nvidia/meta/llama-3.3-70b-instruct) round-trip via the first-slash split;
the native id rides the wire.

NIM exposes raw vLLM behavior and stream_options support varies per model
(some strict vLLM backends 4xx on it), so chat_compat=StrictOpenAi strips
stream_options — streaming works across the whole fleet, usage falls back
to estimation. Snapshot reasoning models carry no effort menus, so kigi
sends no reasoning_effort (which an unsupported strict validator would
400 on). Review: no defects. Logged note: a key lacking the org 'Public
API Endpoints' permission passes /models validation but 403s on chat
(user-fixable edge case).
This commit is contained in:
2026-07-21 14:31:25 -04:00
parent f825132983
commit 9ee40b13d0
4 changed files with 149 additions and 9 deletions
@@ -605,7 +605,8 @@ mod tests {
"google",
"openrouter",
"together",
"cerebras"
"cerebras",
"nvidia"
]
);
assert_eq!(default_id(&built), Some(XAI_API_KEY_METHOD_ID));
@@ -641,7 +642,8 @@ mod tests {
"google",
"openrouter",
"together",
"cerebras"
"cerebras",
"nvidia"
]
);
assert_eq!(default_id(&built), Some(CACHED_TOKEN_AUTH_METHOD_ID));
@@ -670,7 +672,8 @@ mod tests {
"google",
"openrouter",
"together",
"cerebras"
"cerebras",
"nvidia"
]
);
assert_eq!(default_id(&built), Some(CACHED_TOKEN_AUTH_METHOD_ID));
@@ -702,7 +705,8 @@ mod tests {
"google",
"openrouter",
"together",
"cerebras"
"cerebras",
"nvidia"
]
);
assert_eq!(default_id(&built), None);