GPT-OSS 20B vs Qwen3 30B A3B
openai/gpt-oss-20b · qwen/qwen3-30b-a3b
As of October 31, 2026, GPT-OSS 20B is the cheaper API on UberLLM: $0.0000 per million input tokens and $0.0001 per million output, versus $0.0517 / $0.2152 for Qwen3 30B A3B, about 100% less when output is 3 tokens for every input token. These are the median prices buyers paid over the last 24 hours. Both work with one OpenAI-compatible key and UberLLM adds no fee. Choose GPT-OSS 20B to cut cost; choose Qwen3 30B A3B when context size wins. Qwen3 30B A3B carries the larger context window (256,000 vs 128,000 tokens), which can matter for long-document work.
GPT-OSS 20B vs Qwen3 30B A3B, side by side
| Spec | GPT-OSS 20B | Qwen3 30B A3B |
|---|---|---|
| Price · input / 1M | $0.0000 | $0.0517 |
| Price · output / 1M | $0.0001 | $0.2152 |
| Context length | 128,000 | 256,000 |
| Max output tokens | 16,000 | 32,768 |
| Input modalities | text | text |
| Output modalities | text | text |
| Tools / functions | Yes | Yes |
| Reasoning | — | Yes |
| % off list | 99.9% off | 57% off |
Median prices buyers paid on UberLLM over the last 24 hours. No UberLLM fee on requests; adding credit by card costs 5.5% + $0.50. Updated October 31, 2026. A price more than 70% below list is what buyers were billed, not an error. Why so cheap?
GPT-OSS 20B vs Qwen3 30B A3B — FAQ
Is GPT-OSS 20B or Qwen3 30B A3B cheaper?
GPT-OSS 20B is cheaper on a blended token mix, by about 100%. A request costs what its provider charged, UberLLM adds no fee, and every request is listed with its cost in Activity.
What's the price difference between GPT-OSS 20B and Qwen3 30B A3B?
GPT-OSS 20B: $0.0000 in / $0.0001 out per 1M tokens. Qwen3 30B A3B: $0.0517 in / $0.2152 out per 1M tokens. That's roughly a 100% blended difference. Prices are live; UberLLM adds no fee.
Can I use both GPT-OSS 20B and Qwen3 30B A3B with one API?
Yes. Both are served through a single OpenAI-compatible endpoint — set the base URL to https://api.uberllm.dev/v1 and your UberLLM key, then switch between them by changing the model id. Anthropic SDKs work via https://api.uberllm.dev/v1/messages.