Gemini 3.5 Flash vs gpt-oss-20b: pricing comparison
gpt-oss-20b is the cheaper option at $0.02/$0.09 per 1M input/output tokens — Gemini 3.5 Flash ($1.50/$9) costs about 83x more per input token. Full spec-by-spec breakdown below.
| Gemini 3.5 Flash | gpt-oss-20b | |
|---|---|---|
| Input /1M tokens | $1.50 | $0.02 |
| Output /1M tokens | $9 | $0.09 |
| Cache read /1M | $0.15 | $0.01 |
| Context window | 1.0M | 131K |
| Provider | OpenAI | |
| Vision input | Yes | No |
| Released | May 2026 | Aug 2025 |
Real workload costs: Gemini 3.5 Flash vs gpt-oss-20b
Cost per single request at common token profiles — the cheapest model for each workload is highlighted.
| Workload | Tokens (in / out) | Gemini 3.5 Flash | gpt-oss-20b |
|---|---|---|---|
| Chatbot message | 500 / 300 | $0.0034 | <$0.0001 |
| RAG query with context | 4,000 / 500 | $0.01 | $0.0001 |
| Document summarization | 20,000 / 1,000 | $0.04 | $0.0004 |
| Agent coding session | 100,000 / 5,000 | $0.20 | $0.0022 |
Frequently asked questions
Which is cheaper: Gemini 3.5 Flash vs gpt-oss-20b?
gpt-oss-20b is the cheapest on input tokens at $0.02 per 1M, and gpt-oss-20b is cheapest on output at $0.09 per 1M. Gemini 3.5 Flash costs about 83x more per input token than gpt-oss-20b.
How much does Gemini 3.5 Flash cost per 1M tokens?
Gemini 3.5 Flash costs $1.50 per 1M input tokens and $9 per 1M output tokens on the Google API, with a 1.0M-token context window.
How much does gpt-oss-20b cost per 1M tokens?
gpt-oss-20b costs $0.02 per 1M input tokens and $0.09 per 1M output tokens on the OpenAI API, with a 131K-token context window.
What does a typical request cost on Gemini 3.5 Flash vs gpt-oss-20b?
For a typical request with 2,000 input and 500 output tokens: Gemini 3.5 Flash costs $0.0075, gpt-oss-20b costs <$0.0001. At 1,000 requests/day for a month that is $225 for Gemini 3.5 Flash vs $2.43 for gpt-oss-20b.
Estimate your own workload with the per-model calculators (Gemini 3.5 Flash, gpt-oss-20b), browse all current prices on the live model pricing dashboard, or run these models head-to-head on a real prompt (free account required). Prices may differ from provider list prices for batch or tiered usage.
Pricing data via the OpenRouter models API.