Qwen2.5 72B: API Provider Benchmarking & Analysis
Alibaba has launched a newer model, Qwen3 235B, we suggest considering this model instead.
For more information, see Comparison of Qwen3 235B to other models and API provider benchmarks for Qwen3 235B.
Comparison Summary
Together.ai Turbo (86 t/s) and Alibaba Cloud (46 t/s) are the fastest providers of Qwen2.5 72B, followed by 
DeepInfra (33 t/s) & 
Hyperbolic (33 t/s).Latency (TTFT): Together.ai Turbo (0.42s) and DeepInfra (0.50s) have the lowest latency for Qwen2.5 72B, followed by Hyperbolic (0.77s) & Alibaba Cloud (1.09s).Blended Price ($/M tokens): DeepInfra ($0.19) and Hyperbolic ($0.40) are the most cost-effective providers for Qwen2.5 72B, followed by & Together.ai Turbo ($1.20).Input Token Price: DeepInfra ($0.12) and Hyperbolic ($0.40) offer the lowest input token prices for Qwen2.5 72B, followed by & Together.ai Turbo ($1.20).Output Token Price: DeepInfra ($0.39) and Hyperbolic ($0.40) offer the lowest output token prices for Qwen2.5 72B, followed by & Together.ai Turbo ($1.20).Highlights
Pricing
Speed vs. Price: Qwen2.5 72B Providers
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).
Price per token, represented as USD per million Tokens. Price is a blend of Input & Output token prices (3:1 ratio).
Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.
Latency vs. Output Speed: Qwen2.5 72B Providers
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).
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 per token, represented as USD per million Tokens. Price is a blend of Input & Output token prices (3:1 ratio).
Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.
Speed
Measured by Output Speed (tokens per second)
Output Speed: Qwen2.5 72B Providers
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).
Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.
Latency
Measured by Time (seconds) to First Token
Time to First Token: Qwen2.5 72B Providers
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.
Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.
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: Qwen2.5 72B Providers
Price per token, represented as USD per million Tokens. Price is a blend of Input & Output token prices (3:1 ratio).
Seconds to receive a 500 token response considering input processing time, 'thinking' time of reasoning models, and output speed.
Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.
API Features
Function (Tool) Calling & JSON Mode: Qwen2.5 72B Providers
Indicates whether the provider supports function calling in their API. Function calling is also known as 'Tool Calling'.
Indicates whether the provider supports JSON mode in their API. When JSON mode is enabled, the models will always return a valid JSON object.
Context Window: Qwen2.5 72B Providers
Maximum number of combined input & output tokens. Output tokens commonly have a significantly lower limit (varied by model).
While models have their own context window, in cases this is limited by providers.