Microsoft Azure: Models Intelligence, Performance & Price
Analysis of Microsoft Azure'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 Microsoft Azure for your use-case. For more details including relating to our methodology, see our FAQs. Models analyzed: o1, o3-mini, o1-preview, o1-mini, GPT-4o (Aug '24), GPT-4o (May '24), GPT-4o (Nov '24), GPT-4o mini, o3, GPT-4.1 nano, o3-mini (high), GPT-4.1 mini, o4-mini (high), GPT-4.1, Llama 3.3 70B, Llama 3.1 405B, Llama 3.1 70B, Llama 3.1 8B, Llama 4 Maverick (FP8), Llama 4 Scout, Mistral Large 2 (Nov '24), Mistral Large 2 (Jul '24), DeepSeek R1, DeepSeek V3 (Mar' 25), Phi-4, Phi-3 Medium 14B, Phi-4 Mini, Phi-4 Multimodal, Command-R+ (Apr '24), Command-R (Mar '24), Jamba 1.5 Large, Jamba 1.5 Mini, GPT-4 Turbo, Llama 3 70B, Llama 3 8B, Mistral Small (Feb '24), and DeepSeek V3 (Dec '24).
Link:
Azure Model Comparison Summary
Intelligence:
o4-mini (high) and o3 are the highest quality models offered by Azure, followed by o3-mini (high), o3-mini & o1.Output Speed (tokens/s): GPT-4.1 nano (304 t/s) and o1-mini (262 t/s) are the fastest models offered by Azure, followed by 
Llama 3.1 8B, o3-mini (high) & o3-mini.Latency (seconds): Llama 3.1 8B (0.30s) and Phi-4 Multimodal (0.34s) are the lowest latency models offered by Azure, followed by Phi-4 Mini, Llama 4 Maverick (FP8) & Llama 4 Scout.Blended Price ($/M tokens): Phi-4 Mini ($0.00) and Phi-4 Multimodal ($0.00) are the cheapest models offered by Azure, followed by GPT-4.1 nano, Phi-4 & 
Jamba 1.5 Mini.Context Window Size: GPT-4.1 nano (1m) and GPT-4.1 mini (1m) are the largest context window models offered by Azure, followed by GPT-4.1, Jamba 1.5 Large & Jamba 1.5 Mini.
o4-mini (high) and o3 are the highest quality models offered by Azure, followed by o3-mini (high), o3-mini & o1.Output Speed (tokens/s): GPT-4.1 nano (304 t/s) and o1-mini (262 t/s) are the fastest models offered by Azure, followed by Llama 3.1 8B, o3-mini (high) & o3-mini.Latency (seconds): Llama 3.1 8B (0.30s) and Phi-4 Multimodal (0.34s) are the lowest latency models offered by Azure, followed by Phi-4 Mini, Llama 4 Maverick (FP8) & Llama 4 Scout.Blended Price ($/M tokens): Phi-4 Mini ($0.00) and Phi-4 Multimodal ($0.00) are the cheapest models offered by Azure, followed by GPT-4.1 nano, Phi-4 & Jamba 1.5 Mini.Context Window Size: GPT-4.1 nano (1m) and GPT-4.1 mini (1m) are the largest context window models offered by Azure, followed by GPT-4.1, Jamba 1.5 Large & Jamba 1.5 Mini.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 | |||||||||
 | o4-mini (high) | 200k | 70 | $1.93 | 30.5 | 150.15 | 166.55 | N/A | |
| o3 | 128k | 67 | $17.50 | 13.3 | 202.49 | 240.08 | N/A | |
| o3-mini (high) | 200k | 66 | $1.93 | 195.9 | 42.97 | 45.52 | N/A | |
| o3-mini | 200k | 63 | $1.93 | 190.6 | 12.64 | 15.26 | N/A | |
| o1 | 200k | 62 | $26.25 | 107.9 | 28.94 | 33.57 | N/A | |
|  DeepSeek R1 | 128k | 60 | $2.36 | 113.8 | 0.52 | 25.54 | 20.63 | |
| o1-mini | 128k | 54 | $2.12 | 261.7 | 9.80 | 11.71 | N/A | |
| DeepSeek V3 (Mar' 25) | 128k | 53 | $2.00 | 123.5 | 0.56 | 4.61 | N/A | |
| GPT-4.1 mini | 1m | 53 | $0.70 | 140.5 | 0.75 | 4.30 | N/A | |
| GPT-4.1 | 1m | 53 | $3.50 | 66.6 | 1.27 | 8.78 | N/A | |
| Llama 4 Maverick (FP8) | 128k | 51 | $0.61 | 63.4 | 0.34 | 8.23 | N/A | |
| DeepSeek V3 (Dec '24) | 128k | 46 | $2.00 | 66.1 | 0.61 | 8.17 | N/A | |
| Llama 4 Scout | 128k | 43 | $0.34 | 39.0 | 0.35 | 13.17 | N/A | |
| GPT-4o (Nov '24) | 128k | 41 | $4.38 | 124.0 | 1.22 | 5.25 | N/A | |
| Llama 3.3 70B | 128k | 41 | $0.71 | 53.0 | 0.44 | 9.87 | N/A | |
| GPT-4.1 nano | 1m | 41 | $0.17 | 303.6 | 0.53 | 2.18 | N/A | |
| GPT-4o (May '24) | 128k | 41 | $7.50 | 59.2 | 1.18 | 9.63 | N/A | |
| Llama 3.1 405B | 128k | 40 | $8.00 | 31.4 | 0.49 | 16.39 | N/A | |
| Phi-4 | 16k | 40 | $0.22 | 43.6 | 0.49 | 11.95 | N/A | |
|  Mistral Large 2 (Nov '24) | 128k | 38 | $3.00 | 29.7 | 0.52 | 17.34 | N/A | |
| Mistral Large 2 (Jul '24) | 128k | 37 | $3.00 | 36.3 | 0.51 | 14.30 | N/A | |
| GPT-4o mini | 128k | 36 | $0.26 | 153.5 | 0.98 | 4.24 | N/A | |
| Llama 3.1 70B | 128k | 35 | $2.90 | 54.9 | 0.45 | 9.55 | N/A | |
| Jamba 1.5 Large | 256k | 29 | $3.50 | 51.0 | 0.67 | 10.46 | N/A | |
| Llama 3 70B | 8k | 27 | $2.90 | 18.9 | 0.76 | 27.21 | N/A | |
| Phi-4 Multimodal | 128k | 27 | $0.00 | 21.1 | 0.34 | 24.00 | N/A | |
| Phi-4 Mini | 128k | 26 | $0.00 | 56.7 | 0.34 | 9.16 | N/A | |
| Phi-3 Medium 14B | 128k | 25 | $0.30 | 50.6 | 0.44 | 10.32 | N/A | |
| Llama 3.1 8B | 128k | 24 | $0.38 | 226.3 | 0.30 | 2.50 | N/A | |
| Mistral Small (Feb '24) | 33k | 23 | $1.50 | 88.4 | 0.38 | 6.03 | N/A | |
| Llama 3 8B | 8k | 21 | $0.38 | 73.8 | 0.38 | 7.15 | N/A | |
|  Command-R+ (Apr '24) | 128k | 20 | $6.00 | 50.6 | 0.59 | 10.47 | N/A | |
| Jamba 1.5 Mini | 256k | 18 | $0.25 | 82.3 | 0.47 | 6.55 | N/A | |
| Command-R (Mar '24) | 128k | 15 | $0.75 | 80.7 | 0.46 | 6.65 | N/A | |
| o1-preview | 128k | $28.88 | 153.9 | 22.34 | 25.59 | N/A | ||
| GPT-4o (Aug '24) | 128k | $4.38 | 114.9 | 0.76 | 5.11 | N/A | ||
| GPT-4 Turbo | 128k | $15.00 | 49.2 | 1.60 | 11.76 | 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.