Fireworks: Models Intelligence, Performance & Price
Analysis of Fireworks's models across key metrics including quality, price, output speed, latency, context window & more. This analysis is intended to support you in choosing the best model provided by Fireworks for your use-case. For more details including relating to our methodology, see our FAQs. Models analyzed: Llama 3.3 70B, Llama 3.1 405B, Llama 3.2 90B (Vision), Llama 3.1 70B, Llama 3.2 11B (Vision), Llama 3.1 8B, Llama 3.2 3B, Mistral Small 3, Mixtral 8x22B, Mixtral 8x7B, DeepSeek R1 (Fast), DeepSeek V3 (Mar' 25), Qwen2.5 72B, Qwen2.5 Coder 32B, QwQ-32B, Yi-Large, Llama 3 70B, Llama 3 8B, DeepSeek V3, and QwQ 32B-Preview.
Link:
Fireworks Model Comparison Summary
Intelligence:
DeepSeek R1 (Fast) and 
QwQ-32B are the highest quality models offered by Fireworks, followed by DeepSeek V3 (Mar' 25), DeepSeek V3 & QwQ 32B-Preview.Output Speed (tokens/s):
Llama 3.1 8B (209 t/s) and 
Mixtral 8x7B (178 t/s) are the fastest models offered by Fireworks, followed by Llama 3.2 3B, Llama 3.1 70B & Llama 3.3 70B.Latency (seconds): Llama 3.2 3B (0.25s) and Mixtral 8x7B (0.28s) are the lowest latency models offered by Fireworks, followed by Llama 3 8B, Llama 3.2 11B (Vision) & Llama 3.1 8B.Blended Price ($/M tokens): Llama 3.2 3B ($0.10) and Llama 3.2 11B (Vision) ($0.20) are the cheapest models offered by Fireworks, followed by Llama 3.1 8B, Llama 3 8B & Mixtral 8x7B.Context Window Size: DeepSeek R1 (Fast) (164k) and DeepSeek V3 (Mar' 25) (160k) are the largest context window models offered by Fireworks, followed by Qwen2.5 72B, QwQ-32B & Llama 3.3 70B.
DeepSeek R1 (Fast) and QwQ-32B are the highest quality models offered by Fireworks, followed by DeepSeek V3 (Mar' 25), DeepSeek V3 & QwQ 32B-Preview.Output Speed (tokens/s): Llama 3.1 8B (209 t/s) and Mixtral 8x7B (178 t/s) are the fastest models offered by Fireworks, followed by Llama 3.2 3B, Llama 3.1 70B & Llama 3.3 70B.Latency (seconds): Llama 3.2 3B (0.25s) and Mixtral 8x7B (0.28s) are the lowest latency models offered by Fireworks, followed by Llama 3 8B, Llama 3.2 11B (Vision) & Llama 3.1 8B.Blended Price ($/M tokens): Llama 3.2 3B ($0.10) and Llama 3.2 11B (Vision) ($0.20) are the cheapest models offered by Fireworks, followed by Llama 3.1 8B, Llama 3 8B & Mixtral 8x7B.Context Window Size: DeepSeek R1 (Fast) (164k) and DeepSeek V3 (Mar' 25) (160k) are the largest context window models offered by Fireworks, followed by Qwen2.5 72B, QwQ-32B & Llama 3.3 70B.Highlights
Intelligence
Artificial Analysis Intelligence Index; Higher is better
Speed
Output Tokens per Second; Higher is better
Price
USD per 1M Tokens; Lower is better
Parallel Queries:
Prompt Length:
Features | Model Intelligence | Price | Output tokens/s | Latency | |||
|---|---|---|---|---|---|---|---|
Further Analysis | |||||||
 | DeepSeek R1 (Fast) | 164k | 60 | $4.25 | 108.3 | 0.83 | |
| QwQ-32B | 131k | 58 | $0.90 | 131.1 | 0.63 | |
| DeepSeek V3 (Mar' 25) | 160k | 52 | $0.90 | 63.6 | 0.84 | |
| DeepSeek V3 | 128k | 46 | $1.31 | 64.3 | 0.86 | |
| QwQ 32B-Preview | 33k | 43 | $0.90 | 53.1 | 0.52 | |
| Llama 3.3 70B | 128k | 41 | $0.90 | 146.6 | 0.50 | |
| Llama 3.1 405B | 128k | 40 | $3.00 | 84.6 | 0.64 | |
| Qwen2.5 72B | 131k | 40 | $0.90 | 41.1 | 0.43 | |
| Qwen2.5 Coder 32B | 33k | 36 | $0.90 | 58.1 | 0.39 | |
| Llama 3.1 70B | 128k | 35 | $0.90 | 148.4 | 0.45 | |
| Mistral Small 3 | 32k | 35 | $0.90 | 39.7 | 0.55 | |
| Llama 3.2 90B (Vision) | 128k | 33 | $0.90 | 42.2 | 0.40 | |
|  Yi-Large | 32k | 28 | $3.00 | 68.8 | 0.46 | |
| Llama 3 70B | 8k | 27 | $0.90 | 141.8 | 0.37 | |
| Mixtral 8x22B | 65k | 26 | $1.20 | 82.1 | 0.39 | |
| Llama 3.1 8B | 128k | 24 | $0.20 | 208.7 | 0.33 | |
| Llama 3 8B | 8k | 21 | $0.20 | 127.6 | 0.28 | |
| Llama 3.2 3B | 128k | 20 | $0.10 | 166.2 | 0.25 | |
| Mixtral 8x7B | 33k | 17 | $0.50 | 177.7 | 0.28 | |
| Llama 3.2 11B (Vision) | 128k | $0.20 | 106.4 | 0.29 |
Key definitions
Context window: Maximum number of combined input & output tokens. Output tokens commonly have a significantly lower limit (varied by model).
Output Speed: Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).
Latency: Time to first token of tokens received, in seconds, after API request sent. For models which do not support streaming, this represents time to receive the completion.
Price: Price per token, represented as USD per million Tokens. Price is a blend of Input & Output token prices (3:1 ratio).
Output Price: Price per token generated by the model (received from the API), represented as USD per million Tokens.
Input Price: Price per token included in the request/message sent to the API, represented as USD per million Tokens.
Time period: Metrics are 'live' and are based on the past 72 hours of measurements, measurements are taken 8 times a day for single requests and 2 times per day for parallel requests.