LLM APIs services compared: free tiers and pricing
17 llm apis services, each with its free tier and lowest paid price as read from its pricing page.
| Service | Free tier | Lowest paid price | Checked |
|---|---|---|---|
| Amazon Bedrock | AWS Free Tier credits for new accounts | — | 2026-10-23 |
| Anthropic (Claude API) | — | $0.10 per 1M input tokens, $0.50 per 1M output (prompts up to 100K) | 2026-10-22 |
| Azure OpenAI | $200 Azure credit for 30 days (new accounts) | GPT-6.1 Sol (short context, Global) $2 per 1M input, $10 per 1M output | 2026-10-23 |
| Cerebras Inference | Free $5 credit to start | From $0.35 per 1M tokens | 2026-10-23 |
| Cognition | Free use on Pro and above; slow on Free | Included in Devin Pro $20/mo | 2026-10-23 |
| Cohere | Trial API key (rate limited) | Command-light $0.30 per 1M input, $0.60 per 1M output | 2026-10-23 |
| DeepInfra | — | From $0.06 per 1M input tokens and $0.18 per 1M output (smallest listed model) | 2026-10-23 |
| DeepSeek API | — | $0.15 per 1M input (cache miss, off-peak), $0.60 per 1M output | 2026-10-23 |
| Fireworks AI | — | $0.008 per 1M tokens (embedding models up to 150M parameters) | 2026-10-25 |
| Google Gemini API | Free tier (Google AI Studio, data used to improve products) | $0.30 per 1M input tokens, $2.50 per 1M output | 2026-10-23 |
| Groq | — | GPT OSS 20B: $0.075 per 1M input tokens, $0.30 per 1M output | 2026-10-25 |
| Mistral AI | Free API tier for experimentation | Mistral Large $0.50 per 1M input tokens, $1.50 per 1M output | 2026-10-23 |
| OpenAI API | — | $0.10 per 1M input tokens, $0.50 per 1M output | 2026-10-23 |
| OpenRouter | Free: 25+ free models, 4 free providers | Pay-as-you-go: model list price plus a 5.5% fee on credit purchases | 2026-10-23 |
| Perplexity API | — | web_search tool $2.50 per 1,000 calls; embeddings from $0.004 per 1M tokens | 2026-10-23 |
| Together AI | — | From $0.15 per 1M input tokens (e.g. GLM-5.3-Flash, $0.50 output) | 2026-10-23 |
| xAI API | — | $2.00 per 1M input tokens, $6.00 per 1M output | 2026-10-23 |

