← Back to all tools
Visit site
Overall score
3.1/ 5
SME Fit3/5
pricing pattern unclear + free tier
JTBD4/5
solid named JTBD
Integration3/5
API
Trust4/5
mature, founded 2022
Quality1/5
no public rating
Compliance4/5
SOC 2 + GDPR
About
Fireworks AI is a managed inference provider focused on open-source LLMs and image / audio models. Competes with Together AI and Groq on speed and price for Llama, Mistral, DeepSeek, Qwen, and other open-weights families. Strong on fine-tuning workflows.
Best for: Builders running open-source LLMs at scale who want OpenAI-compatible API ergonomics, fast inference, and a serious fine-tuning surface without the operational burden of self-hosting.
Pricing
| Tier | Monthly | Annual /mo | Billing | Notes |
|---|---|---|---|---|
| Pay-as-you-go | n/a | n/a | usage | All hosted models;Per-token serverless inference;Fine-tuning at usage-based pricing;Free $1 starter credit · Per-million-token pricing varies by model. |
| Enterprise | n/a | n/a | custom | Dedicated deployments;Reserved capacity;SOC 2 + HIPAA;Custom data residency;SLAs · Contact sales for pricing. |
Startup offer
Included in hub programs
Key features
- Sub-second inference on popular open models
- OpenAI-compatible API
- Fine-tuning service
- Image and audio model hosting
- Function calling and JSON mode
- Dedicated deployment option
Integrations
OpenAI-compatible APILangChainLlamaIndexHugging FaceReplicate
Trust & compliance
- Stage range
- n/a
- Founded
- 2022
- Status
- active
- SOC 2
- yes
- GDPR
- yes
- Data residency
- us
- External rating
- n/a
- Last verified
- May 2026
Reviews
Be the first to share your experience.