Amazon Bedrock: Models Intelligence, Performance & Price
This analysis is intended to support you in choosing the best model provided by Amazon Bedrock for your use-case.
Most Intelligent
Updated
GPT-6 Astra (max)
Claude Opus 5 (max)Claude Opus 5 (xhigh)Claude Opus 5 (high)GPT-5.6 Sol (max)Intelligence index
Total 93 models
Fastest
Ministral 3 3B
Grok 4.3 (high)Grok 4.3 (medium)Nova MicroGrok 4.3 (low)Output speed
Total 93 models
Lowest Price
Nova Micro
Gemma 3 4BNova Lite
NVIDIA Nemotron Nano 9B V2 (Non-reasoning)gpt-oss-20b (low)Blended price (per 1M tokens)
Total 93 models
Amazon offers 93 models, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across models.
- For intelligence, the top models on Amazon are GPT-6 Astra (max) (53), Claude Opus 5 (max) (51), and Claude Opus 5 (xhigh) (50).
- For output speed, the fastest models are Ministral 3 3B (528 t/s), Grok 4.3 (high) (284 t/s), and Grok 4.3 (medium) (283 t/s). Speed varies significantly across models, with a 104% difference between the fastest and slowest.
- For latency, GPT-5.4 (Non-reasoning) (0.56s), Llama 3.1 8B (0.66s), and GPT-5.5 (Non-reasoning) (0.69s) offer the lowest time to first answer token.
- For pricing, Nova Micro ($0.03), Gemma 3 4B ($0.04), and Nova Lite ($0.05) offer the lowest blended prices per 1M tokens. Prices vary up to 3.4x across models.
- For context window size, Claude Opus 5 (max) (1M), Claude Opus 5 (xhigh) (1M), and Claude Opus 5 (high) (1M) support the largest context windows on Amazon.
Highlights
Intelligence Evaluations
Artificial Analysis Intelligence Index
Intelligence Evaluations
Agentic knowledge work, (Elo-500)/2000
Agentic real-world work tasks, (Elo-500)/2000
Agentic SaaS workflows
Agentic coding & terminal use
Coding
Reasoning & knowledge
Professional document reasoning, All-pass
Physics reasoning
Knowledge
1 - hallucination rate
Long context reasoning
Legal agentic work, criterion pass rate
Agentic business operations
Quantitative analysis on spreadsheets & documents
Instruction following
Agentic tool use
Long-horizon agentic tasks
Kubernetes incident root-cause analysis
Visual reasoning
Intelligence Index vs. Price
Context Window
Context Window
Pricing
Intelligence Index vs. Price
Performance Summary
Output Speed vs. Price
Speed
Measured by Output Speed (tokens per second)
Output Speed
Latency
Measured by Time (seconds) to First Token
Latency: Time To First Answer Token
End-to-End Response Time
Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed
End-to-End Response Time vs. Price
Further Analysis | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
GPT-6 Astra (max) | 524k | Proprietary | 53 | $3.26 | 70 | 405.84 | 413.03 | -- | |||
Claude Opus 5 (max) | 1M | Proprietary | 51 | $5.86 | 51 | 39.50 | 49.27 | -- | |||
Claude Opus 5 (xhigh) | 1M | Proprietary | 50 | $4.88 | 51 | 32.43 | 42.31 | -- | |||
Claude Opus 5 (high) | 1M | Proprietary | 48 | $3.61 | 51 | 12.79 | 22.57 | -- | |||
GPT-5.6 Sol (max) | 1M | Proprietary | 47 | $3.70 | 116 | 74.62 | 78.94 | -- | |||
Claude Opus 5 (medium) | 1M | Proprietary | 45 | $2.19 | 50 | 4.62 | 14.70 | -- | |||
GPT-5.6 Terra (max) | 1M | Proprietary | 42 | $1.95 | 143 | 115.11 | 118.62 | -- | |||
Claude Opus 4.8 (max) | 1M | Proprietary | 42 | $4.02 | 52 | 15.59 | 25.23 | -- | |||
Claude Opus 4.7 (max) | 1M | Proprietary | 41* | -- | 58 | 1.77 | 10.40 | -- | |||
Claude Opus 5 (low) | 1M | Proprietary | 40 | $1.10 | 48 | 2.73 | 13.17 | -- | |||
GPT-5.4 (xhigh) | 272k | Proprietary | 39* | -- | 153 | 76.35 | 79.62 | -- | |||
GPT-5.5 (xhigh) | 272k | Proprietary | 39 | $5.98 | 123 | 29.93 | 33.99 | -- | |||
Claude Sonnet 5 (max) | 1M | Proprietary | 38 | $6.21 | 78 | 183.83 | 190.25 | -- | |||
GPT-5.6 Luna (max) | 1M | Proprietary | 38 | $1.02 | 247 | 84.08 | 86.11 | -- | |||
GPT-5.5 (high) | 272k | Proprietary | 37 | $3.26 | 124 | 11.61 | 15.63 | -- | |||
GPT-5.5 (medium) | 272k | Proprietary | 34 | $1.82 | 124 | 3.02 | 7.04 | -- | |||
Claude Opus 4.6 (max) | 1M | Proprietary | 32* | -- | 45 | 9.17 | 20.36 | -- | |||
Claude Opus 4.7 (Non-reasoning, high) | 1M | Proprietary | 31* | -- | 56 | 1.61 | 10.53 | -- | |||
GPT-5.5 (low) | 272k | Proprietary | 31* | -- | 116 | 1.20 | 5.51 | -- | |||
Claude Sonnet 4.6 (max) | 200k | Proprietary | 30 | $2.83 | 75 | 66.77 | 73.44 | -- | |||
Claude Opus 4.5 | 200k | Proprietary | 29* | -- | 46 | 13.35 | 24.14 | -- | |||
Claude Sonnet 5 (Non-reasoning) | 1M | Proprietary | 29* | -- | 57 | 1.98 | 10.79 | -- | |||
GPT-5.6 Sol (Non-reasoning) | 1M | Proprietary | 28* | -- | 111 | 0.86 | 5.38 | -- | |||
GPT-5.4 (low) | 272k | Proprietary | 28* | -- | 133 | 1.13 | 4.89 | -- | |||
Claude Opus 4.6 (Non-reasoning, high) | 1M | Proprietary | 26* | -- | 43 | 1.77 | 13.42 | -- | |||
Grok 4.3 (high) | 524k | Proprietary | 25 | $0.35 | 284 | 8.07 | 9.83 | -- | |||
Grok 4.3 (medium) | 524k | Proprietary | 25* | -- | 283 | 3.98 | 5.75 | -- | |||
Claude Sonnet 4.6 (Non-reasoning, high) | 200k | Proprietary | 25* | -- | 54 | 1.39 | 10.65 | -- | |||
Grok 4.3 (low) | 524k | Proprietary | 24* | -- | 259 | 2.72 | 4.65 | -- | |||
Claude Opus 4.5 (Non-reasoning) | 200k | Proprietary | 24* | -- | 44 | 1.89 | 13.22 | -- | |||
 Kimi K2.5 | 256k | Open | 23* | -- | 93 | 1.32 | 38.75 | 32.04 | |||
GPT-5.5 (Non-reasoning) | 272k | Proprietary | 23* | -- | 118 | 0.69 | 4.93 | -- | |||
Claude 4.1 Opus | 200k | Proprietary | 23* | -- | -- | -- | -- | -- | |||
GPT-5.6 Terra (Non-reasoning) | 1M | Proprietary | 22* | -- | 124 | 0.75 | 4.78 | -- | |||
 GLM-4.7 | 200k | Open | 22* | -- | 96 | 1.34 | 27.26 | 20.74 | |||
Kimi K2 Thinking | 256k | Open | 22* | -- | 121 | 1.39 | 21.99 | 16.48 | |||
 DeepSeek V3.2 | 128k | Open | 21* | -- | 89 | 1.41 | 29.57 | 22.54 | |||
Claude 4.5 Sonnet | 1M | Proprietary | 21 | $0.56 | 48 | 7.63 | 18.14 | -- | |||
Claude 4.5 Sonnet (Non-reasoning) | 1M | Proprietary | 19* | -- | 45 | 1.53 | 12.62 | -- | |||
 MiniMax-M2 | 205k | Open | 19* | -- | 106 | 1.08 | 24.60 | 18.82 | |||
Claude 4.1 Opus (Non-reasoning) | 200k | Proprietary | 19* | -- | -- | -- | -- | -- | |||
GPT-5.4 (Non-reasoning) | 272k | Proprietary | 18* | -- | 124 | 0.56 | 4.60 | -- | |||
Claude 4.5 Haiku | 200k | Proprietary | 18 | $0.20 | 97 | 12.18 | 17.36 | -- | |||
GLM-4.7 (Non-reasoning) | 200k | Open | 17* | -- | 101 | 1.41 | 6.37 | -- | |||
GPT-5.6 Luna (Non-reasoning) | 1M | Proprietary | 17* | -- | 211 | 0.82 | 3.19 | -- | |||
DeepSeek V3.2 (Non-reasoning) | 128k | Open | 16* | -- | 59 | 1.62 | 10.12 | -- | |||
Claude 4.5 Haiku (Non-reasoning) | 200k | Proprietary | 15* | -- | 99 | 0.86 | 5.92 | -- | |||
GLM-4.7-Flash | 200k | Open | 15* | -- | 205 | 0.98 | 13.17 | 9.76 | |||
Grok 4.3 (Non-reasoning) | 524k | Proprietary | 15 | $0.28 | 242 | 0.73 | 2.80 | -- | |||
Nova 2.0 Pro Preview (medium) | 256k | Proprietary | 14* | -- | 112 | 13.71 | 36.02 | 17.85 | |||
DeepSeek V3.1 (Non-reasoning) | 128k | Open | 14* | -- | 149 | 1.38 | 4.73 | -- | |||
Nova 2.0 Omni (medium) | 1M | Proprietary | 14* | -- | -- | -- | -- | -- | |||
Nova 2.0 Omni (medium) | 1M | Proprietary | 14* | -- | -- | -- | -- | -- | |||
DeepSeek V3.1 | 128k | Open | 13* | -- | 140 | 1.37 | 19.26 | 14.31 | |||
Nova 2.0 Lite (high) | 1M | Proprietary | 13* | -- | -- | -- | -- | -- | |||
Nova 2.0 Lite (high) | 1M | Proprietary | 13* | -- | 157 | 14.07 | 29.99 | 12.74 | |||
Nova 2.0 Pro Preview (low) | 256k | Proprietary | 13* | -- | 112 | 10.07 | 32.31 | 17.79 | |||
Nova 2.0 Lite (medium) | 1M | Proprietary | 12* | -- | 152 | 10.71 | 27.14 | 13.14 | |||
gpt-oss-120b (high) | 131k | Open | 12 | $0.11 | 66 | 0.94 | 38.87 | 30.34 | |||
 Qwen3 235B 2507 (Non-reasoning) | 256k | Open | 12* | -- | 93 | 1.26 | 6.61 | -- | |||
Qwen3 Coder 480B | 262k | Open | 12* | -- | 96 | 1.20 | 6.43 | -- | |||
Nova 2.0 Lite (low) | 1M | Proprietary | 12* | -- | 165 | 7.23 | 22.38 | 12.12 | |||
DeepSeek R1 (Jan) | 128k | Open | 11 | $0.22 | 174 | 1.05 | 17.39 | 13.47 | |||
Nova 2.0 Omni (low) | 1M | Proprietary | 11* | -- | -- | -- | -- | -- | |||
GLM-4.7-Flash (Non-reasoning) | 200k | Open | 11* | -- | 208 | 0.99 | 3.39 | -- | |||
gpt-oss-120b (low) | 131k | Open | 10* | -- | 77 | 0.94 | 33.43 | 25.99 | |||
Qwen3 Coder Next | 128k | Open | 10 | $0.78 | 63 | 1.03 | 9.00 | -- | |||
Nova 2.0 Pro Preview (Non-reasoning) | 256k | Proprietary | 10* | -- | 104 | 1.04 | 5.85 | -- | |||
gpt-oss-20b (low) | 131k | Open | 10* | -- | 128 | 9.95 | 29.43 | 15.59 | |||
Mistral Large 3 | 256k | Open | 10 | $0.10 | 198 | 1.09 | 3.61 | -- | |||
Qwen3 Coder 30B A3B | 262k | Open | 10* | -- | 233 | 0.95 | 3.09 | -- | |||
 Llama 4 Maverick | 128k | Open | 9 | $0.04 | 218 | 0.75 | 3.04 | -- | |||
Nova Premier | 1M | Proprietary | 9* | -- | 33 | 2.81 | 18.00 | -- | |||
gpt-oss-20b (high) | 131k | Open | 9 | $0.01 | 152 | 32.79 | 49.22 | 13.14 | |||
Nova 2.0 Lite (Non-reasoning) | 1M | Proprietary | 9* | -- | 153 | 1.03 | 4.29 | -- | |||
Magistral Small 1.2 | 128k | Open | 9* | -- | 74 | 1.58 | 35.56 | 27.18 | |||
Nova 2.0 Omni (Non-reasoning) | 1M | Proprietary | 8* | -- | -- | -- | -- | -- | |||
Claude 3.5 Sonnet (Oct) | 200k | Proprietary | 8* | -- | -- | -- | -- | -- | |||
Llama 3.3 70B | 128k | Open | 8* | -- | 141 | 1.12 | 4.66 | -- | |||
Qwen3 32B (Non-reasoning) | 32.8k | Open | 7* | -- | 115 | 1.12 | 5.46 | -- | |||
Claude 3.5 Sonnet (June) | 200k | Proprietary | 7* | -- | -- | -- | -- | -- | |||
Nova Pro | 300k | Proprietary | 7* | -- | 131 | 1.13 | 4.94 | -- | |||
Llama 3.1 8B | 128k | Open | 7* | -- | 174 | 0.66 | 3.53 | -- | |||
NVIDIA Nemotron Nano 9B V2 (Non-reasoning) | 131k | Open | 7* | -- | 166 | 1.10 | 4.11 | -- | |||
Mistral Large 2 (Jul) | 128k | Open | 7* | -- | 31 | 1.77 | 18.00 | -- | |||
Nova Lite | 300k | Proprietary | 7* | -- | 172 | 0.93 | 3.83 | -- | |||
Llama 3.1 70B Standard | 128k | Open | 7* | -- | 97 | 1.31 | 6.47 | -- | |||
Llama 3.1 70B Latency Optimized | 128k | Open | 7* | -- | 109 | 1.24 | 5.84 | -- | |||
Llama 4 Scout | 128k | Open | 6 | $0.11 | 162 | 0.82 | 3.91 | -- | |||
Ministral 3 14B | 256k | Open | 6 | $0.14 | 242 | 0.92 | 2.98 | -- | |||
 Jamba 1.5 Large | 256k | Open | 6* | -- | 47 | 2.71 | 13.24 | -- | |||
Nova Micro | 130k | Proprietary | 6* | -- | 269 | 0.86 | 2.72 | -- | |||
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning) | 128k | Open | 6* | -- | 200 | 1.07 | 3.57 | -- | |||
Mistral Large (Feb) | 32.8k | Proprietary | 6* | -- | 43 | 1.51 | 13.16 | -- | |||
Claude 3 Haiku | 200k | Proprietary | 6* | -- | 86 | 0.85 | 6.67 | -- | |||
Ministral 3 8B | 256k | Open | 5 | $0.07 | 235 | 1.00 | 3.13 | -- | |||
Llama 3 70B | 8.19k | Open | 5* | -- | -- | -- | -- | -- | |||
Jamba 1.5 Mini | 256k | Open | 5* | -- | 61 | 4.47 | 12.68 | -- | |||
Mixtral 8x7B | 32.8k | Open | 5* | -- | 24 | 1.11 | 21.72 | -- | |||
Mistral 7B | 8.19k | Open | 5* | -- | -- | -- | -- | -- | |||
Gemma 3 27B | 128k | Open | 5 | $0.47 | 64 | 1.35 | 9.22 | -- | |||
Ministral 3 3B | 256k | Open | 5 | $0.04 | 528 | 0.87 | 1.82 | -- | |||
Gemma 3 4B | 128k | Open | 5* | -- | 198 | 0.97 | 3.49 | -- | |||
Llama 3 8B | 8.19k | Open | 5* | -- | -- | -- | -- | -- | |||
Gemma 3 12B | 128k | Open | 4 | $0.24 | 96 | 1.31 | 6.52 | -- | |||
Key definitions
Frequently Asked Questions
Common questions about Amazon Bedrock
Amazon Bedrock offers 93 models that we track: GPT-6 Astra (max), Claude Opus 5 (max), Claude Opus 5 (xhigh), Claude Opus 5 (high), GPT-5.6 Sol (max), Claude Opus 5 (medium), GPT-5.6 Terra (max), Claude Opus 4.8 (max), Claude Opus 4.7 (max), Claude Opus 5 (low), GPT-5.4 (xhigh), GPT-5.5 (xhigh), Claude Sonnet 5 (max), GPT-5.6 Luna (max), GPT-5.5 (high), GPT-5.5 (medium), Claude Opus 4.6 (max), Claude Opus 4.7 (Non-reasoning, high), GPT-5.5 (low), Claude Sonnet 4.6 (max), Claude Opus 4.5, Claude Sonnet 5 (Non-reasoning), GPT-5.6 Sol (Non-reasoning), GPT-5.4 (low), Claude Opus 4.6 (Non-reasoning, high), Grok 4.3 (high), Grok 4.3 (medium), Claude Sonnet 4.6 (Non-reasoning, high), Grok 4.3 (low), Claude Opus 4.5 (Non-reasoning), Kimi K2.5, GPT-5.5 (Non-reasoning), GLM-4.7, Kimi K2 Thinking, DeepSeek V3.2, GPT-5.6 Terra (Non-reasoning), Claude 4.5 Sonnet, Claude 4.5 Sonnet (Non-reasoning), MiniMax-M2, GPT-5.4 (Non-reasoning), Claude 4.5 Haiku, GLM-4.7 (Non-reasoning), GPT-5.6 Luna (Non-reasoning), DeepSeek V3.2 (Non-reasoning), Claude 4.5 Haiku (Non-reasoning), GLM-4.7-Flash, Grok 4.3 (Non-reasoning), Nova 2.0 Pro Preview (medium), DeepSeek V3.1 (Non-reasoning), DeepSeek V3.1, Nova 2.0 Lite (high), Nova 2.0 Pro Preview (low), Nova 2.0 Lite (medium), gpt-oss-120b (high), Qwen3 235B 2507 (Non-reasoning), Qwen3 Coder 480B, Nova 2.0 Lite (low), DeepSeek R1 (Jan), GLM-4.7-Flash (Non-reasoning), gpt-oss-120b (low), Qwen3 Coder Next, Nova 2.0 Pro Preview (Non-reasoning), gpt-oss-20b (low), Mistral Large 3, Qwen3 Coder 30B A3B, Llama 4 Maverick, Nova Premier, gpt-oss-20b (high), Nova 2.0 Lite (Non-reasoning), Magistral Small 1.2, Llama 3.3 70B, Qwen3 32B (Non-reasoning), Nova Pro, Llama 3.1 8B, NVIDIA Nemotron Nano 9B V2 (Non-reasoning), Mistral Large 2 (Jul), Nova Lite, Llama 3.1 70B Standard, Llama 3.1 70B Latency Optimized, Llama 4 Scout, Ministral 3 14B, Jamba 1.5 Large, Nova Micro, NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning), Mistral Large (Feb), Claude 3 Haiku, Ministral 3 8B, Jamba 1.5 Mini, Mixtral 8x7B, Gemma 3 27B, Ministral 3 3B, Gemma 3 4B, and Gemma 3 12B.
The most intelligent model available on Amazon Bedrock is GPT-6 Astra (max) with an Intelligence Index score of 53.
The fastest model on Amazon Bedrock by output speed is Ministral 3 3B at 527.7 tokens per second.
The model with the lowest time to first answer token on Amazon Bedrock is GPT-5.4 (Non-reasoning) at 0.56s. Lower latency means faster initial response time.
The most affordable model on Amazon Bedrock by blended price is Nova Micro at $0.03 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on Amazon Bedrock vary up to 284x across models, from $0.03 per 1M tokens for Nova Micro to $7.70 per 1M tokens for GPT-6 Astra (max).
23 of 93 models on Amazon Bedrock support JSON mode for structured output.
91 of 93 models on Amazon Bedrock support function calling (tool use).
Yes, Amazon Bedrock offers 44 reasoning models: GPT-6 Astra (max), Claude Opus 5 (max), Claude Opus 5 (xhigh), Claude Opus 5 (high), GPT-5.6 Sol (max), Claude Opus 5 (medium), GPT-5.6 Terra (max), Claude Opus 4.8 (max), Claude Opus 4.7 (max), Claude Opus 5 (low), GPT-5.4 (xhigh), GPT-5.5 (xhigh), Claude Sonnet 5 (max), GPT-5.6 Luna (max), GPT-5.5 (high), GPT-5.5 (medium), Claude Opus 4.6 (max), GPT-5.5 (low), Claude Sonnet 4.6 (max), Claude Opus 4.5, GPT-5.4 (low), Grok 4.3 (high), Grok 4.3 (medium), Grok 4.3 (low), Kimi K2.5, GLM-4.7, Kimi K2 Thinking, DeepSeek V3.2, Claude 4.5 Sonnet, MiniMax-M2, Claude 4.5 Haiku, GLM-4.7-Flash, Nova 2.0 Pro Preview (medium), DeepSeek V3.1, Nova 2.0 Lite (high), Nova 2.0 Pro Preview (low), Nova 2.0 Lite (medium), gpt-oss-120b (high), Nova 2.0 Lite (low), DeepSeek R1 (Jan), gpt-oss-120b (low), gpt-oss-20b (low), gpt-oss-20b (high), and Magistral Small 1.2. Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, 41 of 93 models on Amazon Bedrock are open weight models: Kimi K2.5, GLM-4.7, Kimi K2 Thinking, DeepSeek V3.2, MiniMax-M2, GLM-4.7 (Non-reasoning), DeepSeek V3.2 (Non-reasoning), GLM-4.7-Flash, DeepSeek V3.1 (Non-reasoning), DeepSeek V3.1, gpt-oss-120b (high), Qwen3 235B 2507 (Non-reasoning), Qwen3 Coder 480B, DeepSeek R1 (Jan), GLM-4.7-Flash (Non-reasoning), gpt-oss-120b (low), Qwen3 Coder Next, gpt-oss-20b (low), Mistral Large 3, Qwen3 Coder 30B A3B, Llama 4 Maverick, gpt-oss-20b (high), Magistral Small 1.2, Llama 3.3 70B, Qwen3 32B (Non-reasoning), Llama 3.1 8B, NVIDIA Nemotron Nano 9B V2 (Non-reasoning), Mistral Large 2 (Jul), Llama 3.1 70B Standard, Llama 3.1 70B Latency Optimized, Llama 4 Scout, Ministral 3 14B, Jamba 1.5 Large, NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning), Ministral 3 8B, Jamba 1.5 Mini, Mixtral 8x7B, Gemma 3 27B, Ministral 3 3B, Gemma 3 4B, and Gemma 3 12B.
Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.
When choosing a model on Amazon Bedrock, consider: intelligence (for quality-sensitive tasks), output speed (for throughput-intensive tasks), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and features like context window size, JSON mode, or function calling support.