diff --git a/.github/workflows/update-readme.yml b/.github/workflows/update-readme.yml index 6cc2dd3..b3e4389 100644 --- a/.github/workflows/update-readme.yml +++ b/.github/workflows/update-readme.yml @@ -37,6 +37,7 @@ jobs: GOOGLE_APPLICATION_CREDENTIALS: ${{ steps.auth.outputs.credentials_file_path }} LAMBDA_API_KEY: ${{ secrets.LAMBDA_API_KEY }} MISTRAL_API_KEY: ${{ secrets.MISTRAL_API_KEY }} + SCALEWAY_API_KEY: ${{ secrets.SCALEWAY_API_KEY }} run: | python -u src/pull_available_models.py - name: Remove credentials diff --git a/README.md b/README.md index ac8e165..a7e533e 100644 --- a/README.md +++ b/README.md @@ -41,7 +41,7 @@ This lists various services that provide free access or credits towards API-base Mixtral 8x7B14,400 requests/day
5,000 tokens/minute Whisper Large v37,200 audio-seconds/minute
2,000 requests/day Whisper Large v3 Turbo7,200 audio-seconds/minute
2,000 requests/day -OpenRouter20 requests/minute
200 requests/dayGemini 2.0 Flash Experimental +OpenRouter20 requests/minute
200 requests/dayGemini 2.0 Flash Experimental Gemini Experimental 1114 Gemini Experimental 1206 Gemma 2 9B Instruct @@ -61,7 +61,6 @@ This lists various services that provide free access or credits towards API-base Qwen 2 7B Instruct Toppy M 7B Zephyr 7B Beta -google/gemini-2.0-flash-thinking-exp:free Google AI Studio Data is used for training (when used outside of the UK/CH/EEA/EU). @@ -178,6 +177,13 @@ This lists various services that provide free access or credits towards API-base Mistral Nemo 240712 requests/minute Mixtral 8x22B Instruct12 requests/minute Mixtral 8x7B Instruct12 requests/minute +Scaleway Generative APIs (Free Beta)BGE-Multilingual-Gemma2600 requests/minute
1,000,000 tokens/minute +Llama 3.1 70B Instruct300 requests/minute
100,000 tokens/minute +Llama 3.1 8B Instruct300 requests/minute
100,000 tokens/minute +Mistral Nemo 2407300 requests/minute
100,000 tokens/minute +Pixtral 12B (2409)300 requests/minute
100,000 tokens/minute +Qwen2.5 Coder 32B Instruct +sentence-t5-xxl600 requests/minute
1,000,000 tokens/minute Cloudflare Workers AI10,000 tokens/dayDeepseek Coder 6.7B Base (AWQ) Deepseek Coder 6.7B Instruct (AWQ) Deepseek Math 7B Instruct @@ -321,6 +327,7 @@ This lists various services that provide free access or credits towards API-base Llama 3.1 405B Base Llama 3.1 405B Base (FP8) Llama 3.1 405B Instruct +Llama 3.1 405B Instruct Virtuals Llama 3.1 70B Instruct Llama 3.1 8B Instruct Llama 3.2 3B Instruct @@ -331,7 +338,6 @@ This lists various services that provide free access or credits towards API-base Qwen2-VL 7B Instruct Qwen2.5 72B Instruct Qwen2.5 Coder 32B Instruct -meta-llama/meta-llama-3.1-405b-instruct-virtuals AI21 $10 for 3 months diff --git a/src/pull_available_models.py b/src/pull_available_models.py index f961960..ef5ab36 100644 --- a/src/pull_available_models.py +++ b/src/pull_available_models.py @@ -3,6 +3,7 @@ from collections import defaultdict import logging import json +from bs4 import BeautifulSoup import requests import os from dotenv import load_dotenv @@ -159,6 +160,13 @@ MODEL_TO_NAME_MAPPING = { "llama-3.3-70b-specdec": "Llama 3.3 70B (Speculative Decoding)", "@cf/meta/llama-3.3-70b-instruct-fp8-fast": "Llama 3.3 70B Instruct (FP8)", "google/gemini-2.0-flash-exp:free": "Gemini 2.0 Flash Experimental", + "qwen2.5-coder-32b-instruct": "Qwen2.5 Coder 32B Instruct", + "bge-multilingual-gemma2": "BGE-Multilingual-Gemma2", + "pixtral-12b-2409": "Pixtral 12B (2409)", + "google/gemini-2.0-flash-thinking-exp:free": "Gemini 2.0 Flash Thinking Experimental", + "sentence-t5-xxl": "sentence-t5-xxl", + "meta-llama/meta-llama-3.1-405b-instruct-virtuals": "Llama 3.1 405B Instruct Virtuals", + "llama-3.1-8b-instruct": "Llama 3.1 8B Instruct", } @@ -184,6 +192,7 @@ OPENROUTER_IGNORED_MODELS = { "google/gemini-exp-1114:free" "google/gemini-exp-1206:free" "google/gemini-2.0-flash-exp:free", + "google/gemini-2.0-flash-thinking-exp:free" } # Ignore gemini experimental free models because rate limits mean they are unusable. @@ -293,7 +302,7 @@ def fetch_openrouter_models(logger): continue if ":free" not in model["id"]: continue - if model["id"] in OPENROUTER_IGNORED_MODELS: + if model["id"].lower() in OPENROUTER_IGNORED_MODELS: logger.debug(f"Ignoring model {model['id']}") continue ret_models.append( @@ -637,6 +646,77 @@ ONLY OUTPUT JSON!""", return ret_models +def fetch_scaleway_models(logger): + logger.info("Fetching Scaleway models...") + r = requests.get( + "https://api.scaleway.ai/v1/models", + headers={"Authorization": f"Bearer {os.environ['SCALEWAY_API_KEY']}"}, + ) + r.raise_for_status() + models = r.json()["data"] + logger.info(f"Fetched {len(models)} models from Scaleway") + ret_models = [] + for model in models: + ret_models.append( + { + "id": model["id"], + "name": get_model_name(model["id"]), + } + ) + ret_models = sorted(ret_models, key=lambda x: x["name"]) + logger.info("Fetching Scaleway rate limits") + r = requests.get( + "https://www.scaleway.com/en/docs/ai-data/generative-apis/reference-content/rate-limits/" + ) + r.raise_for_status() + # Extract
content + soup = BeautifulSoup(r.text, "html.parser") + body = str(soup.find("main")) + prompt = f""" +Here is the web page to extract data from: +```html +{body} +``` + """ + client = Mistral(api_key=os.environ["MISTRAL_API_KEY"]) + logger.info("Extracting model rate limits from the provided web page...") + chat_response = client.chat.complete( + model="mistral-large-latest", + messages=[ + { + "role": "system", + "content": """Extract the model rate limits as only integers from the provided web page into JSON of the following format: +```json +{ + "model name here": { + "requests/minute": 10, + "tokens/minute": 100 + } +} +``` +ONLY OUTPUT JSON!""", + }, + { + "role": "user", + "content": prompt, + }, + ], + response_format={ + "type": "json_object", + }, + temperature=0, + ) + logger.info(chat_response.choices[0].message.content) + extracted_data = json.loads(chat_response.choices[0].message.content) + for model in extracted_data: + for m in ret_models: + if m["id"] == model: + m["limits"] = extracted_data[model] + break + + return ret_models + + def get_human_limits(model): if "limits" not in model: return "" @@ -655,6 +735,7 @@ def main(): hyperbolic_logger = create_logger("Hyperbolic") # lambda_logger = create_logger("Lambda Labs") samba_logger = create_logger("SambaNova") + scaleway_logger = create_logger("Scaleway") gemini_models = fetch_gemini_limits(google_ai_studio_logger) openrouter_models = fetch_openrouter_models(openrouter_logger) @@ -664,6 +745,7 @@ def main(): github_models = fetch_github_models(github_logger) # lambda_models = fetch_lambda_models(lambda_logger) samba_models = fetch_samba_models(samba_logger) + scaleway_models = fetch_scaleway_models(scaleway_logger) groq_models = fetch_groq_models(groq_logger) table = """ @@ -824,6 +906,17 @@ def main(): table += f"" table += "\n" + for idx, model in enumerate(scaleway_models): + table += "" + if idx == 0: + table += '" + table += '' + table += f"" + table += f"" + table += "\n" + for idx, model in enumerate(cloudflare_models): table += "" if idx == 0: diff --git a/src/requirements.txt b/src/requirements.txt index 08fd926..532c7e0 100644 --- a/src/requirements.txt +++ b/src/requirements.txt @@ -2,3 +2,4 @@ requests python-dotenv google-cloud-quotas mistralai +beautifulsoup4 \ No newline at end of file
{get_human_limits(model)}
' + table += 'Scaleway Generative APIs (Free Beta)' + table += "{model['name']}{get_human_limits(model)}