Skip to content

Commit 6cdc11e

Browse files
authored
update the correct naming for llama 3 70b instruct models (#5074)
1 parent 35df351 commit 6cdc11e

File tree

3 files changed

+39
-22
lines changed

3 files changed

+39
-22
lines changed

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

Lines changed: 24 additions & 18 deletions
Original file line numberDiff line numberDiff line change
@@ -1804,77 +1804,77 @@ exports[`Registry Snapshots endpoint configurations snapshot 1`] = `
18041804
"*",
18051805
],
18061806
},
1807-
"llama-3.3-70b-instruct:groq": {
1807+
"llama-3.3-70b-instruct:novita": {
18081808
"context": 131072,
18091809
"crossRegion": false,
1810-
"maxTokens": 32678,
1811-
"modelId": "llama-3.3-70b-versatile",
1810+
"maxTokens": 120000,
1811+
"modelId": "meta-llama/llama-3.3-70b-instruct",
18121812
"parameters": [
18131813
"frequency_penalty",
1814-
"logit_bias",
18151814
"max_tokens",
18161815
"min_p",
18171816
"presence_penalty",
18181817
"repetition_penalty",
1818+
"seed",
18191819
"stop",
18201820
"temperature",
18211821
"tool_choice",
18221822
"tools",
18231823
"top_k",
18241824
"top_p",
18251825
],
1826-
"provider": "groq",
1826+
"provider": "novita",
18271827
"ptbEnabled": true,
18281828
"regions": [
18291829
"*",
18301830
],
18311831
},
1832-
"llama-3.3-70b-instruct:novita": {
1832+
"llama-3.3-70b-instruct:openrouter": {
18331833
"context": 131072,
18341834
"crossRegion": false,
1835-
"maxTokens": 120000,
1835+
"maxTokens": 8192,
18361836
"modelId": "meta-llama/llama-3.3-70b-instruct",
18371837
"parameters": [
18381838
"frequency_penalty",
18391839
"max_tokens",
1840-
"min_p",
18411840
"presence_penalty",
18421841
"repetition_penalty",
1842+
"response_format",
18431843
"seed",
18441844
"stop",
18451845
"temperature",
18461846
"tool_choice",
18471847
"tools",
18481848
"top_k",
1849+
"top_logprobs",
18491850
"top_p",
18501851
],
1851-
"provider": "novita",
1852+
"provider": "openrouter",
18521853
"ptbEnabled": true,
18531854
"regions": [
18541855
"*",
18551856
],
18561857
},
1857-
"llama-3.3-70b-instruct:openrouter": {
1858+
"llama-3.3-70b-versatile:groq": {
18581859
"context": 131072,
18591860
"crossRegion": false,
1860-
"maxTokens": 8192,
1861-
"modelId": "meta-llama/llama-3.3-70b-instruct",
1861+
"maxTokens": 32678,
1862+
"modelId": "llama-3.3-70b-versatile",
18621863
"parameters": [
18631864
"frequency_penalty",
1865+
"logit_bias",
18641866
"max_tokens",
1867+
"min_p",
18651868
"presence_penalty",
18661869
"repetition_penalty",
1867-
"response_format",
1868-
"seed",
18691870
"stop",
18701871
"temperature",
18711872
"tool_choice",
18721873
"tools",
18731874
"top_k",
1874-
"top_logprobs",
18751875
"top_p",
18761876
],
1877-
"provider": "openrouter",
1877+
"provider": "groq",
18781878
"ptbEnabled": true,
18791879
"regions": [
18801880
"*",
@@ -5238,11 +5238,16 @@ exports[`Registry Snapshots verify registry state 1`] = `
52385238
{
52395239
"model": "llama-3.3-70b-instruct",
52405240
"providers": [
5241-
"groq",
52425241
"novita",
52435242
"openrouter",
52445243
],
52455244
},
5245+
{
5246+
"model": "llama-3.3-70b-versatile",
5247+
"providers": [
5248+
"groq",
5249+
],
5250+
},
52465251
{
52475252
"model": "llama-4-maverick",
52485253
"providers": [
@@ -5479,6 +5484,7 @@ exports[`Registry Snapshots verify registry state 1`] = `
54795484
"llama-3.1-8b-instruct",
54805485
"llama-3.1-8b-instruct-turbo",
54815486
"llama-3.3-70b-instruct",
5487+
"llama-3.3-70b-versatile",
54825488
"llama-4-maverick",
54835489
"llama-4-scout",
54845490
"llama-guard-4",
@@ -5510,7 +5516,7 @@ exports[`Registry Snapshots verify registry state 1`] = `
55105516
"totalArchivedConfigs": 0,
55115517
"totalEndpoints": 152,
55125518
"totalModelProviderConfigs": 152,
5513-
"totalModelsWithPtb": 66,
5519+
"totalModelsWithPtb": 67,
55145520
"totalProviders": 14,
55155521
}
55165522
`;

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

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -181,7 +181,7 @@ export const endpoints = {
181181
"*": {},
182182
},
183183
},
184-
"llama-3.3-70b-instruct:groq": {
184+
"llama-3.3-70b-versatile:groq": {
185185
providerModelId: "llama-3.3-70b-versatile",
186186
provider: "groq",
187187
author: "meta-llama",

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

Lines changed: 14 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -34,14 +34,25 @@ export const models = {
3434
modality: { inputs: ["text", "image"], outputs: ["text"] },
3535
tokenizer: "GPT",
3636
},
37-
"llama-3.3-70b-instruct": {
37+
"llama-3.3-70b-versatile": {
3838
name: "Meta Llama 3.3 70B Versatile",
3939
author: "meta-llama",
4040
description:
41-
"Flagship 70B instruction-tuned model for high-quality chat, coding, and reasoning with strong instruction-following and multilingual support.",
41+
"Llama-3.3-70B-Versatile is Meta's advanced multilingual large language model, optimized for a wide range of natural language processing tasks. With 70 billion parameters, it offers high performance across various benchmarks while maintaining efficiency suitable for diverse applications.",
4242
contextLength: 131_072,
4343
maxOutputTokens: 32_678,
44-
created: "2024-12-01T00:00:00.000Z",
44+
created: "2024-12-06T00:00:00.000Z",
45+
modality: { inputs: ["text"], outputs: ["text"] },
46+
tokenizer: "GPT",
47+
},
48+
"llama-3.3-70b-instruct": {
49+
name: "Meta Llama 3.3 70B Instruct",
50+
author: "meta-llama",
51+
description:
52+
"The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model is optimized for multilingual dialogue use cases and outperforms many of the available open source and closed chat models on common industry benchmarks. Supported languages: English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.",
53+
contextLength: 128_000,
54+
maxOutputTokens: 16_400,
55+
created: "2024-12-06T00:00:00.000Z",
4556
modality: { inputs: ["text"], outputs: ["text"] },
4657
tokenizer: "GPT",
4758
},

0 commit comments

Comments
 (0)