A comprehensive guide to the APIs of major Large Language Models.
Base URL: https://api.openai.com/v1
Authentication: Bearer Token in Authorization header.
August 2025 Update: OpenAI has introduced GPT-5 and new open-weight models. GPT-4o is expected to be phased out for API users in favor of GPT-5.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| gpt-5, gpt-5-mini, gpt-5-nano | /chat/completions | Text, Images, Advanced Tool Use |
| gpt-oss-120b, gpt-oss-20b | /chat/completions | Text (Open-weight reasoning models) |
| gpt-image-1 | /images/generations | Text-to-Image |
| TTS (Text-to-Speech) | /audio/speech | Text-to-Audio |
| Whisper | /audio/transcriptions | Audio-to-Text |
Base URL: https://api.anthropic.com/v1
Authentication: API Key via `x-api-key` header.
August 2025 Update: Anthropic released Claude Opus 4.1 on Aug 5, 2025, with improved coding and reasoning.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| claude-opus-4-1-20250805 (Claude Opus 4.1) | /messages | Text, Images |
| claude-opus-4, claude-sonnet-4, etc. | /messages | Text, Images (Older versions) |
Base URL: https://generativelanguage.googleapis.com/v1beta
Authentication: API Key passed as a `key` query parameter.
August 2025 Update: Gemini 2.5 models are now the primary recommendation. Gemini 1.5 models are no longer available for new projects.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| gemini-2.5-pro | /models/gemini-2.5-pro:generateContent | Text, Images, Audio, Video |
| gemini-2.5-flash | /models/gemini-2.5-flash:generateContent | Text, Images, Audio, Video |
| Veo (Video Generation) | /models/veo-3.0-generate-preview:predict | Image/Text-to-Video |
| gemini-embedding-001 | /models/gemini-embedding-001:embedContent | Text |
Base URL: https://api.mistral.ai/v1
Authentication: Bearer Token in Authorization header.
August 2025 Update: Mistral announced Codestral 25.08, a new version of their coding model with significant performance improvements.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| mistral-large-latest, mistral-small-latest, etc. | /chat/completions | Text |
| codestral-2508 | /chat/completions | Code |
| mistral-embed | /embeddings | Text |
Base URL: https://api.x.ai/v1
Authentication: Bearer Token in Authorization header.
August 2025 Update: Elon Musk announced that Grok 2 will be open-sourced soon. An official API release for Grok 2 is anticipated to follow.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| grok-1.5-vision-preview, grok-1.5 | /chat/completions | Text, Images |
Base URL: https://api.deepseek.com
Authentication: Bearer Token in Authorization header.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| deepseek-chat, deepseek-coder | /chat/completions | Text |
Base URL: https://dashscope.aliyuncs.com/api/v1
Authentication: API Key in Authorization header (e.g., `Authorization: Bearer YOUR_API_KEY`).
August 2025 Update: The powerful `Qwen3 Coder 480B` model was announced, available through partners like Cerebras.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| qwen-max, qwen-vl-max, etc. | /services/aigc/text-generation/generation | Text, Image, Video, Audio |
Base URL: https://api.llama.com/v1
Authentication: Bearer Token in Authorization header.
Future Update: Meta has announced the Llama 4 family of models, which will be faster and more capable. Keep an eye out for their official API release.
| Model Name | API Endpoint | Modalities |
|---|---|---|
| Llama 3 | /chat/completions | Text |