integrate nebius llama 3.1-8b-turbo (#5023)

* integrate nebius llama 3.1-8b-fast

* remove unneeded usage processor

Author: juliettech13

packages/__tests__/cost/__snapshots__/registrySnapshots.test.ts.snap

@@ -1459,6 +1459,35 @@ exports[`Registry Snapshots endpoint configurations snapshot 1`] = `
         "*",
       ],
     },
+    "llama-3.1-8b-instruct-turbo:nebius": {
+      "context": 128000,
+      "crossRegion": false,
+      "maxTokens": 8192,
+      "modelId": "meta-llama/Meta-Llama-3.1-8B-Instruct-fast",
+      "parameters": [
+        "frequency_penalty",
+        "functions",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p",
+      ],
+      "provider": "nebius",
+      "ptbEnabled": true,
+      "regions": [
+        "*",
+      ],
+    },
     "llama-3.1-8b-instruct:deepinfra": {
       "context": 131072,
       "crossRegion": false,
@@ -3366,6 +3395,7 @@ exports[`Registry Snapshots model coverage snapshot 1`] = `
     "groq",
     "groq",
     "groq",
+    "nebius",
     "novita",
     "novita",
     "novita",
@@ -4107,6 +4137,13 @@ exports[`Registry Snapshots pricing snapshot 1`] = `
         "web_search": 0,
       },
     ],
+    "nebius": [
+      {
+        "input": 3e-8,
+        "output": 9e-8,
+        "threshold": 0,
+      },
+    ],
     "novita": [
       {
         "input": 1.3e-7,
@@ -4688,6 +4725,7 @@ exports[`Registry Snapshots verify registry state 1`] = `
       "model": "llama-3.1-8b-instruct-turbo",
       "providers": [
         "deepinfra",
+        "nebius",
       ],
     },
     {
@@ -4838,6 +4876,10 @@ exports[`Registry Snapshots verify registry state 1`] = `
       "modelCount": 14,
       "provider": "groq",
     },
+    {
+      "modelCount": 1,
+      "provider": "nebius",
+    },
     {
       "modelCount": 11,
       "provider": "novita",
@@ -4928,9 +4970,9 @@ exports[`Registry Snapshots verify registry state 1`] = `
     "claude-3.5-haiku:openrouter:*",
   ],
   "totalArchivedConfigs": 0,
-  "totalEndpoints": 137,
-  "totalModelProviderConfigs": 137,
+  "totalEndpoints": 138,
+  "totalModelProviderConfigs": 138,
   "totalModelsWithPtb": 58,
-  "totalProviders": 12,
+  "totalProviders": 13,
 }
 `;

packages/cost/models/authors/meta/llama/endpoints.ts

@@ -101,7 +101,7 @@ export const endpoints = {
       "logit_bias",
       "functions",
       "tools",
-      "tool_choice"
+      "tool_choice",
     ],
     ptbEnabled: true,
     endpointConfigs: {
@@ -529,6 +529,43 @@ export const endpoints = {
       "*": {},
     },
   },
+  "llama-3.1-8b-instruct-turbo:nebius": {
+    providerModelId: "meta-llama/Meta-Llama-3.1-8B-Instruct-fast",
+    provider: "nebius",
+    author: "meta-llama",
+    pricing: [
+      {
+        threshold: 0,
+        input: 0.00000003, // $0.03/1M tokens
+        output: 0.00000009, // $0.09/1M tokens
+      },
+    ],
+    quantization: "fp8",
+    contextLength: 128_000,
+    maxCompletionTokens: 8_192,
+    supportedParameters: [
+      "structured_outputs",
+      "response_format",
+      "max_tokens",
+      "temperature",
+      "top_p",
+      "stop",
+      "frequency_penalty",
+      "presence_penalty",
+      "seed",
+      "top_k",
+      "logit_bias",
+      "logprobs",
+      "top_logprobs",
+      "functions",
+      "tool_choice",
+      "tools",
+    ],
+    ptbEnabled: true,
+    endpointConfigs: {
+      "*": {},
+    },
+  },
   "llama-3.1-8b-instruct:novita": {
     providerModelId: "meta-llama/llama-3.1-8b-instruct",
     provider: "novita",

packages/cost/models/provider-helpers.ts

@@ -51,6 +51,8 @@ export function heliconeProviderToModelProviderName(
       return "deepinfra";
     case "NOVITA":
       return "novita";
+    case "NEBIUS":
+      return "nebius";
     // new registry does not have
     case "LOCAL":
     case "HELICONE":
@@ -66,7 +68,6 @@ export function heliconeProviderToModelProviderName(
     case "QSTASH":
     case "FIRECRAWL":
     case "AVIAN":
-    case "NEBIUS":
     case "OPENPIPE":
     case "CHUTES":
     case "LLAMA":

packages/cost/models/providers/index.ts

@@ -12,6 +12,7 @@ import { GoogleProvider } from "./google";
 import { DeepInfraProvider } from "./deepinfra";
 import { OpenRouterProvider } from "./openrouter";
 import { NovitaProvider } from "./novita";
+import { NebiusProvider } from "./nebius";
 
 // Create singleton instances (stateless, so safe to share)
 export const providers = {
@@ -29,6 +30,7 @@ export const providers = {
   "google-ai-studio": new GoogleProvider(),
   openrouter: new OpenRouterProvider(),
   novita: new NovitaProvider(),
+  nebius: new NebiusProvider()
 } as const;
 
 export type ModelProviderName = keyof typeof providers;

packages/cost/models/providers/nebius.ts

@@ -0,0 +1,17 @@
+import { BaseProvider } from "./base";
+import type { Endpoint, RequestParams } from "../types";
+
+export class NebiusProvider extends BaseProvider {
+  readonly displayName = "Nebius";
+  readonly baseUrl = "https://api.studio.nebius.com/v1/";
+  readonly auth = "api-key" as const;
+  readonly pricingPages = [
+    "https://nebius.com/prices-ai-studio",
+    "https://nebius.com/prices",
+  ];
+  readonly modelPages = ["https://studio.nebius.com/"];
+
+  buildUrl(endpoint: Endpoint, requestParams: RequestParams): string {
+    return `${this.baseUrl}chat/completions`;
+  }
+}

packages/cost/usage/getUsageProcessor.ts

@@ -17,6 +17,7 @@ export function getUsageProcessor(
     case "openai":
     case "azure":
     case "deepinfra":
+    case "nebius":
     case "novita":
       return new OpenAIUsageProcessor();
     case "anthropic":

web/data/providers.ts

@@ -224,6 +224,16 @@ export const providers: Provider[] = [
     apiKeyPlaceholder: "...",
     relevanceScore: 10,
   },
+  {
+    id: "nebius",
+    name: "Nebius",
+    logoUrl: "/assets/home/providers/nebius.webp",
+    description: "Configure your Nebius API keys",
+    docsUrl: "https://docs.helicone.ai/getting-started/integration-methods",
+    apiKeyLabel: "Nebius API Key",
+    apiKeyPlaceholder: "...",
+    relevanceScore: 5,
+  },
 ];
 
 // Mock recently used providers - in a real app, this would come from user data