Fireworks AI review

Fireworks AI ranks fifth and has the lowest starting price we found, $0.05 per 1M input tokens on its smallest highlighted model. Its pricing page explains a useful idea: every serverless request can pick a mode. Standard is the default, Priority holds up at peak load, Fast targets 100 or more tokens per second, Batch is 50% off, and Spot is 30% off on best-effort capacity that can drop to zero. Cached input is billed at half price or less with no write fee. The highlighted catalog includes Kimi K3 at $3 input, GLM 5.3 at $1.40, DeepSeek V4.1 Flash at $0.30 and GPT OSS 120B at $0.15. For steady traffic, Reserved Throughput adds an SLA and 99.9% uptime terms. The catches are that serverless runs on prepaid credits and US-only processing adds a 50% premium on select models.

Who it is for

Teams serving open-weight models in production who want to trade speed against cost per request.

Price

Pay per token; serverless from $0.05 per 1M input tokens.

Serverless Pay per token · From $0.05 / 1M input tokens
Quoted from the vendor pricing page on November 7, 2026.

Key features

• Five serverless modes per request: Standard, Priority, Fast, Batch and Spot
• Batch at 50% off and Spot at 30% off Standard on select models
• Cached input billed at half price or less, with no write fee
• Reserved Throughput with a throughput SLA and 99.9% uptime terms
• Models including Kimi K3, GLM 5.3, DeepSeek V4.1 Flash and OpenAI GPT OSS 120B

Strengths and weaknesses

• Strength: The lowest published starting price on this list at $0.05 per 1M input tokens
• Strength: Switch speed and price mode per request on the same endpoint
• Strength: No cache write fee
• Weakness: Spot capacity can drop to zero, so it needs retries or fallback
• Weakness: US-only serverless costs a 50% premium on select models
• Weakness: Usage draws down prepaid credits

Key facts

• Starting price: $0.05 per 1M input tokens (NVIDIA Nemotron 3.5 Lightning 30B A3B)
• Fast mode: 100+ tokens per second for interactive apps
• Batch discount: 50% off input and output

Where it ranks

Fireworks AI is No. 4 in our Best LLM APIs verdict.