Amazon Bedrock: Models Intelligence, Performance & Price
Analysis of Amazon Bedrock'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 Amazon Bedrock 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 Standard, Llama 3.1 405B Latency Optimized, Llama 3.2 90B (Vision), Llama 3.1 70B Standard, Llama 3.1 70B Latency Optimized, Llama 3.2 11B (Vision), Llama 3.1 8B, Llama 3.2 3B, Llama 3.2 1B, Claude 3.5 Sonnet (Oct), Claude 3.5 Sonnet (June), Claude 3 Opus, Claude 3.5 Haiku Standard, Claude 3.5 Haiku Latency Optimized, Claude 3 Haiku, Claude 3.7 Sonnet, Mistral Large 2 (Jul '24), Mixtral 8x7B, DeepSeek R1, Nova Pro, Nova Pro Latency Optimized, Nova Lite, Nova Micro, Command-R+, Command-R+ (Apr '24), Command-R (Mar '24), Command-R, Llama 3 70B, Llama 3 8B, Claude 3 Sonnet, Claude 2.1, and Mistral Large (Feb '24).
Link:
Amazon Model Comparison Summary
Intelligence:
DeepSeek R1 and 
Claude 3.7 Sonnet are the highest quality models offered by Amazon, followed by Claude 3.5 Sonnet (Oct), 
Llama 3.3 70B & Llama 3.1 405B Standard.Output Speed (tokens/s): Nova Micro (324 t/s) and Nova Lite (278 t/s) are the fastest models offered by Amazon, followed by Llama 3.2 11B (Vision), Llama 3.3 70B & Llama 3.2 1B.Latency (seconds): Nova Micro (0.32s) and Llama 3 8B (0.32s) are the lowest latency models offered by Amazon, followed by 
Mixtral 8x7B, Nova Lite & Nova Pro.Blended Price ($/M tokens): Nova Micro ($0.06) and Llama 3.2 1B ($0.10) are the cheapest models offered by Amazon, followed by Nova Lite, Llama 3.2 3B & Llama 3.2 11B (Vision).Context Window Size: Nova Pro (300k) and Nova Pro Latency Optimized (300k) are the largest context window models offered by Amazon, followed by Nova Lite, Claude 3.5 Sonnet (Oct) & Claude 3.5 Sonnet (June).
DeepSeek R1 and Claude 3.7 Sonnet are the highest quality models offered by Amazon, followed by Claude 3.5 Sonnet (Oct), Llama 3.3 70B & Llama 3.1 405B Standard.Output Speed (tokens/s): Nova Micro (324 t/s) and Nova Lite (278 t/s) are the fastest models offered by Amazon, followed by Llama 3.2 11B (Vision), Llama 3.3 70B & Llama 3.2 1B.Latency (seconds): Nova Micro (0.32s) and Llama 3 8B (0.32s) are the lowest latency models offered by Amazon, followed by Mixtral 8x7B, Nova Lite & Nova Pro.Blended Price ($/M tokens): Nova Micro ($0.06) and Llama 3.2 1B ($0.10) are the cheapest models offered by Amazon, followed by Nova Lite, Llama 3.2 3B & Llama 3.2 11B (Vision).Context Window Size: Nova Pro (300k) and Nova Pro Latency Optimized (300k) are the largest context window models offered by Amazon, followed by Nova Lite, Claude 3.5 Sonnet (Oct) & Claude 3.5 Sonnet (June).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 | End-to-End Response Time | ||||
|---|---|---|---|---|---|---|---|---|---|
Further Analysis | |||||||||
 | DeepSeek R1 | 128k | 60 | $2.36 | 69.3 | 0.44 | 41.53 | 33.88 | |
| Claude 3.7 Sonnet | 200k | 48 | $6.00 | 45.4 | 1.14 | 12.15 | N/A | |
| Claude 3.5 Sonnet (Oct) | 200k | 44 | $6.00 | 47.7 | 0.96 | 11.45 | N/A | |
| Llama 3.3 70B | 128k | 41 | $0.71 | 138.4 | 0.59 | 4.21 | N/A | |
| Llama 3.1 405B Standard | 128k | 40 | $2.40 | 30.4 | 1.83 | 18.29 | N/A | |
| Llama 3.1 405B Latency Optimized | 128k | 40 | $3.00 | 30.0 | 2.13 | 18.81 | N/A | |
| Nova Pro | 300k | 37 | $1.40 | 108.0 | 0.35 | 4.98 | N/A | |
| Nova Pro Latency Optimized | 300k | 37 | $1.75 | 110.3 | 0.39 | 4.93 | N/A | |
| Mistral Large 2 (Jul '24) | 128k | 37 | $3.00 | 33.3 | 0.46 | 15.47 | N/A | |
| Llama 3.1 70B Standard | 128k | 35 | $0.72 | 31.6 | 0.64 | 16.48 | N/A | |
| Llama 3.1 70B Latency Optimized | 128k | 35 | $0.90 | 31.7 | 0.82 | 16.60 | N/A | |
| Claude 3 Opus | 200k | 35 | $30.00 | 25.3 | 1.24 | 21.01 | N/A | |
| Claude 3.5 Haiku Standard | 200k | 35 | $1.60 | 56.2 | 1.30 | 10.20 | N/A | |
| Claude 3.5 Haiku Latency Optimized | 200k | 35 | $2.00 | 59.6 | 1.35 | 9.73 | N/A | |
| Llama 3.2 90B (Vision) | 128k | 33 | $0.72 | 60.6 | 0.52 | 8.77 | N/A | |
| Nova Lite | 300k | 33 | $0.10 | 277.5 | 0.34 | 2.14 | N/A | |
| Nova Micro | 130k | 28 | $0.06 | 324.2 | 0.32 | 1.86 | N/A | |
| Claude 3 Sonnet | 200k | 28 | $6.00 | 52.6 | 0.83 | 10.33 | N/A | |
| Llama 3 70B | 8k | 27 | $2.86 | 47.5 | 0.40 | 10.93 | N/A | |
| Mistral Large (Feb '24) | 33k | 26 | $6.00 | 44.0 | 0.39 | 11.75 | N/A | |
| Claude 2.1 | 200k | 24 | $12.00 | 29.3 | 1.62 | 18.67 | N/A | |
| Llama 3.1 8B | 128k | 24 | $0.22 | 91.5 | 0.37 | 5.83 | N/A | |
|  Command-R+ | 128k | 21 | $6.00 | 47.3 | 0.48 | 11.05 | N/A | |
| Llama 3 8B | 8k | 21 | $0.38 | 103.6 | 0.32 | 5.15 | N/A | |
| Command-R+ (Apr '24) | 128k | 20 | $6.00 | 47.5 | 0.49 | 11.01 | N/A | |
| Llama 3.2 3B | 128k | 20 | $0.15 | 72.2 | 0.47 | 7.39 | N/A | |
| Mixtral 8x7B | 33k | 17 | $0.51 | 79.9 | 0.33 | 6.58 | N/A | |
| Command-R | 128k | 15 | $0.75 | 108.2 | 0.35 | 4.97 | N/A | |
| Command-R (Mar '24) | 128k | 15 | $0.75 | 107.8 | 0.35 | 4.99 | N/A | |
| Llama 3.2 1B | 128k | 10 | $0.10 | 118.0 | 0.46 | 4.70 | N/A | |
| Llama 3.2 11B (Vision) | 128k | $0.16 | 143.7 | 0.48 | 3.96 | N/A | ||
| Claude 3.5 Sonnet (June) | 200k | $6.00 | 47.6 | 0.90 | 11.40 | N/A | ||
| Claude 3 Haiku | 200k | $0.50 | 112.8 | 0.67 | 5.10 | N/A | ||
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): Time to first token received, in seconds, after API request sent. For reasoning models which share reasoning tokens, this will be the first reasoning token. 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.