Qwen3 32B: API Provider Benchmarking & Analysis
Alibaba has launched a newer model, Qwen3 30B A3B 2507, we suggest considering this model instead.
For more information, see Comparison of Qwen3 30B A3B 2507 to other models and API provider benchmarks for Qwen3 30B A3B 2507.
Comparison Summary
Cerebras (2141 t/s) and 
Groq (497 t/s) are the fastest providers of Qwen3 32B, followed by 
SambaNova (286 t/s), 
Amazon (274 t/s) & 
Nebius Fast (127 t/s).Latency (TTFT): Groq (0.21s) and Cerebras (0.22s) have the lowest latency for Qwen3 32B, followed by Nebius Fast (0.58s), Nebius Base (0.64s) & Amazon (0.83s).Blended Price ($/M tokens): Nebius Base ($0.15) and 
DeepInfra (FP8) ($0.15) are the most cost-effective providers for Qwen3 32B, followed by 
Novita (FP8) ($0.19), Amazon ($0.26) & Nebius Fast ($0.30).Input Token Price: Novita (FP8) ($0.10) and Nebius Base ($0.10) offer the lowest input token prices for Qwen3 32B, followed by DeepInfra (FP8) ($0.10), Amazon ($0.15) & Nebius Fast ($0.20).Output Token Price: Nebius Base ($0.30) and DeepInfra (FP8) ($0.30) offer the lowest output token prices for Qwen3 32B, followed by Novita (FP8) ($0.45), Groq ($0.59) & Nebius Fast ($0.60).Highlights
Pricing
Pricing: Input and Output Prices: Qwen3 32B Providers
Price per token included in the request/message sent to the API, represented as USD per million Tokens.
Price per token generated by the model (received from the API), represented as USD per million Tokens.
Speed vs. Price: Qwen3 32B 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.
Speed
Measured by Output Speed (tokens per second)
Output Speed: Qwen3 32B 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 vs. Output Speed: Qwen3 32B 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.
Latency
Measured by Time (seconds) to First Token
Time to First Token: Qwen3 32B 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: Qwen3 32B Providers
Seconds to receive a 500 token response. Key components:
- Input time: Time to receive the first response token
- Thinking time (only for reasoning models): Time reasoning models spend outputting tokens to reason prior to providing an answer. Amount of tokens based on the average reasoning tokens across a diverse set of 60 prompts (methodology details).
- Answer time: Time to generate 500 output tokens, based on output speed
For fair comparison, the number of reasoning tokens is standardized across all providers for each model based on the model's representative query token counts.
Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.
API Features
Function (Tool) Calling & JSON Mode: Qwen3 32B 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: Qwen3 32B 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.