add deferred tool loading for kimi-k3
This commit is contained in:
@@ -1890,6 +1890,23 @@ export const VERCEL_AI_GATEWAY_MODELS = {
|
||||
contextWindow: 262144,
|
||||
maxTokens: 32768,
|
||||
} satisfies Model<"anthropic-messages">,
|
||||
"moonshotai/kimi-k3": {
|
||||
id: "moonshotai/kimi-k3",
|
||||
name: "Kimi K3",
|
||||
api: "anthropic-messages",
|
||||
provider: "vercel-ai-gateway",
|
||||
baseUrl: "https://ai-gateway.vercel.sh",
|
||||
reasoning: true,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
input: 3,
|
||||
output: 15,
|
||||
cacheRead: 0.3,
|
||||
cacheWrite: 0,
|
||||
},
|
||||
contextWindow: 1000000,
|
||||
maxTokens: 256000,
|
||||
} satisfies Model<"anthropic-messages">,
|
||||
"nvidia/nemotron-3-nano-30b-a3b": {
|
||||
id: "nvidia/nemotron-3-nano-30b-a3b",
|
||||
name: "Nemotron 3 Nano 30B A3B",
|
||||
|
||||
Reference in New Issue
Block a user