Qwen3 32B vs Step-3.5-Flash: Specs & Benchmark Comparison
Qwen3 32B is developed by Alibaba, while Step-3.5-Flash comes from StepFun. Qwen3 32B was released in April 2025, and Step-3.5-Flash followed 10 months later in February 2026. Step-3.5-Flash is the larger model at roughly 196 billion parameters, against 33 billion for Qwen3 32B.
We have 9 benchmark results for Qwen3 32B and 7 benchmark results for Step-3.5-Flash, but they overlap on a single test: AIME 2025, where Step-3.5-Flash scores 97.0% against 72.9%.
Step-3.5-Flash is the cheaper API at $0.10 per million input tokens and $0.40 per million output tokens, roughly 4 times cheaper than Qwen3 32B at $0.40 and $0.80. Qwen3 32B takes the larger context window at 128K tokens, compared with 66K for Step-3.5-Flash. Qwen3 32B accepts text as input, while Step-3.5-Flash accepts text, images, audio, and video.
| Characteristic | Qwen3 32B | Step-3.5-Flash |
|---|---|---|
| Company | Alibaba | StepFun |
| Release Date | April 29, 2025 | February 1, 2026 |
| Parameters | 33B | 196B |
| Multimodal | No | Yes |
| Context (input) | 128K | 66K |
| Context (output) | 128K | 8K |
| Input Price / 1M | $0.40 | $0.10 |
| Output Price / 1M | $0.80 | $0.40 |
| Average Score | 72.0% | 78.6% |
| Benchmarks | ||
| AIME 2025 | 72.9% | 97.0% |
Visual Benchmark Comparison
Verdict
Step-3.5-Flash leads in 2 out of 4 comparison categories.
Both models show comparable average scores: Qwen3 32B — 0.7, Step-3.5-Flash — 0.8.
Step-3.5-Flash is 2.4x cheaper: input $0.10/1M vs $0.40/1M tokens.
Qwen3 32B supports a larger context: 128K vs 66K tokens.
Step-3.5-Flash is newer: released 2/1/2026 vs 4/29/2025.
More About These Models
Frequently Asked Questions
Which is better for coding — Qwen3 32B or Step-3.5-Flash?
Which model is cheaper — Qwen3 32B or Step-3.5-Flash?
Which has a larger context window — Qwen3 32B or Step-3.5-Flash?
The Qwen3 32B and Step-3.5-Flash comparison is updated for 2026. Data includes benchmark results, API pricing, context window size and other specifications. For more detailed information, visit the Qwen3 32B or Step-3.5-Flash page. See also the complete list of AI model comparisons.