Qwen3 235B (Reasoning): API Provider Benchmarking & Analysis
Analysis of API providers for Qwen3 235B A22B (Reasoning) across performance metrics including latency (time to first token), output speed (output tokens per second), price and others. API providers benchmarked include Together.ai, Fireworks, Deepinfra, Nebius, Parasail, Alibaba Cloud, and Novita.
Alibaba has launched a newer model, Qwen3 235B 2507 (Reasoning). We suggest considering this model instead of Qwen3 235B (Reasoning). See the following pages for a comparison of Qwen3 235B 2507 (Reasoning) to other models and Qwen3 235B 2507 (Reasoning) API provider benchmarks.
Comparison Summary
Output Speed (tokens/s):
Fireworks (84 t/s) and 
Nebius Base (43 t/s) are the fastest providers of Qwen3 235B (Reasoning), followed by 
Novita (FP8) (38 t/s), 
Deepinfra (FP8) (35 t/s) & 
Together.ai (FP8) (34 t/s).Latency (TTFT): Together.ai (FP8) (0.30s) and Deepinfra (FP8) (0.57s) have the lowest latency for Qwen3 235B (Reasoning), followed by Fireworks (0.59s), Nebius Base (0.66s) & Novita (FP8) (0.81s).Blended Price ($/M tokens): Deepinfra (FP8) ($0.25) and are the most cost-effective providers for Qwen3 235B (Reasoning), followed by .
Fireworks (84 t/s) and Nebius Base (43 t/s) are the fastest providers of Qwen3 235B (Reasoning), followed by Novita (FP8) (38 t/s), Deepinfra (FP8) (35 t/s) & Together.ai (FP8) (34 t/s).Latency (TTFT): Together.ai (FP8) (0.30s) and Deepinfra (FP8) (0.57s) have the lowest latency for Qwen3 235B (Reasoning), followed by Fireworks (0.59s), Nebius Base (0.66s) & Novita (FP8) (0.81s).Blended Price ($/M tokens): Deepinfra (FP8) ($0.25) and are the most cost-effective providers for Qwen3 235B (Reasoning), followed by .Highlights
Intelligence
Artificial Analysis Intelligence Index; Higher is better
