Fireworks AI API pricing, checked at the source
Serverless prices for open models on Fireworks AI, read off Fireworks' own pricing docs with a link on every row. Estimate a monthly bill below, with caching and batch.
Fireworks AI API prices per 1M tokens
Standard prices in USD. Each row links to Fireworks AI's pricing page and shows the day we checked it.
| Model | Input | Cached input | Output | Batch inference (in / out) | Notes | Source |
|---|---|---|---|---|---|---|
| Ember-1 | $3.00 | $0.30 | $15.00 | $1.50 / $7.50 | Sourcechecked 2026-10-25 | |
| Kimi K3 | $3.00 | $0.30 | $15.00 | $1.50 / $7.50 | Sourcechecked 2026-10-25 | |
| Qwen 3.8 Max | $2.00 | $0.25 | $6.00 | $1.00 / $3.00 | Sourcechecked 2026-10-25 | |
| GLM 5.3 | $1.40 | $0.26 | $4.40 | $0.70 / $2.20 | Sourcechecked 2026-10-25 | |
| Muse Spark Open | $1.25 | $0.15 | $4.25 | $0.625 / $2.125 | Runs on US-only capacity; the US-only premium is waived. | Sourcechecked 2026-10-25 |
| MiniMax M3 | $0.30 | $0.06 | $1.20 | $0.15 / $0.60 | Sourcechecked 2026-10-25 | |
| DeepSeek V4.1 Flash | $0.30 | $0.006 | $1.20 | $0.15 / $0.60 | Sourcechecked 2026-10-25 | |
| GLM 5.3 Flash | $0.15 | $0.03 | $0.50 | $0.075 / $0.25 | Sourcechecked 2026-10-25 | |
| OpenAI GPT OSS 120B | $0.15 | $0.015 | $0.60 | $0.075 / $0.30 | Sourcechecked 2026-10-25 | |
| NVIDIA Nemotron 3.5 Lightning 30B A3B | $0.05 | $0.01 | $0.20 | $0.025 / $0.10 | Sourcechecked 2026-10-25 |
Terms on this page: Token · Prompt caching · Context window · Batch API
Estimate your Fireworks AI bill
Enter your monthly tokens and how they split between new input, cached input and output. The result updates as you type.
Write 40k, 250M or 5B. Input, cached input and output tokens together.
100M tokens a month costs from $3.43 (NVIDIA Nemotron 3.5 Lightning 30B A3B) to $203.40 (Kimi K3).
| Model | Per month | Per 1M tokens at this split |
|---|---|---|
| NVIDIA Nemotron 3.5 Lightning 30B A3B | $3.43 | $0.0343 |
| OpenAI GPT OSS 120B | $9.42 | $0.0942 |
| GLM 5.3 Flash | $9.79 | $0.0979 |
| DeepSeek V4.1 Flash | $17.45 | $0.17448 |
| MiniMax M3 | $20.58 | $0.2058 |
| Muse Spark Open | $76.20 | $0.762 |
| GLM 5.3 | $88.88 | $0.8888 |
| Qwen 3.8 Max | $118.50 | $1.185 |
| Ember-1 | $203.40 | $2.034 |
| Kimi K3 | $203.40 | $2.034 |
Standard prices in USD on 2026-10-25, before tax. Long-prompt tiers are left out; see the price table.
Fireworks AI embeddings prices
Prices per 1M input tokens. Embedding models have no output price.
| Model | Price as listed | Per 1M tokens | Notes | Source |
|---|---|---|---|---|
| Embedding models up to 150M parameters | $0.008 per 1M tokens | $0.008 | Sourcechecked 2026-10-25 | |
| Embedding models 150M to 350M parameters | $0.016 per 1M tokens | $0.016 | Sourcechecked 2026-10-25 | |
| Qwen3 8B | $0.10 per 1M tokens | $0.10 | Sourcechecked 2026-10-25 |
Write 40k, 250M or 5B.
100M tokens a month costs from $0.80 (Embedding models up to 150M parameters) to $10.00 (Qwen3 8B).
| Model | Per month | Per 1M tokens |
|---|---|---|
| Embedding models up to 150M parameters | $0.80 | $0.008 |
| Embedding models 150M to 350M parameters | $1.60 | $0.016 |
| Qwen3 8B | $10.00 | $0.10 |
List prices in USD on 2026-10-25, before tax, free tiers and add-ons. Rows billed another way are in the price table only.
How the bill is worked out
Batch is half price
Cached input costs less
Prepaid credits
Size based prices for other models
Fireworks AI API pricing questions
How much does the Fireworks AI API cost?
Fireworks AI bills its language models per million tokens, with separate prices for input, cached input and output. On 2026-10-25, standard prices ran from $0.05 (NVIDIA Nemotron 3.5 Lightning 30B A3B) to $3.00 (Kimi K3) per 1M input tokens, and $0.20 (NVIDIA Nemotron 3.5 Lightning 30B A3B) to $15.00 (Kimi K3) per 1M output tokens. Embeddings runs from $0.008 (Embedding models up to 150M parameters) to $0.10 (Qwen3 8B) per 1M tokens.
Does Fireworks AI offer a batch discount?
Yes. Batch inference costs 50% less than the standard price. It suits work that does not need an answer straight away, such as overnight document runs.
How does prompt caching change a Fireworks AI bill?
Every request is billed on input, cached input and output tokens. Cached input tokens come from the prompt cache and are priced lower.
Where do these prices come from?
From Fireworks AI's own pricing page, https://docs.fireworks.ai/serverless/pricing, checked on 2026-10-25. Every row links to it. When a price cannot be confirmed, the row says to check the vendor's page instead of showing a number.
How does Fireworks price models it does not list by name?
By size. Fireworks prices other models by parameter count, from $0.10 per 1M tokens under 4B parameters up to $0.90 over 16B.
Compare and plan
Building with the Fireworks AI API? OceanLab builds AI into business workflows, with people kept in charge. AI automation →
