GPT-5.4 Nano vs Qwen3.8 Flash: cost per task
GPT-5.4 Nano is cheaper on 10 of 14 tasks and about 1.4x cheaper on average (geometric mean).
| Task | GPT-5.4 Nano | Qwen3.8 Flash | Difference |
|---|---|---|---|
| Answer a chat question | 0.15¢ | 0.08¢ | 1.9x |
| Reply to a customer email | 0.02¢ | 0.04¢ | 2.2x |
| Write a 1,500-word blog post | 0.30¢ | 0.75¢ | 2.5x |
| Write 10 product descriptions | 0.13¢ | 0.76¢ | 5.7x |
| Translate 1,000 words | 0.20¢ | 0.08¢ | 2.5x |
| Proofread one page (500 words) | 0.11¢ | 0.10¢ | 1.1x |
| Summarize a 20-page PDF | 0.33¢ | 0.32¢ | 1.0x |
| Extract data from an invoice | 0.08¢ | 0.11¢ | 1.4x |
| Analyze a 200-row spreadsheet | 0.44¢ | 0.85¢ | 1.9x |
| Answer from 5 retrieved documents (RAG) | 0.12¢ | 0.19¢ | 1.7x |
| Classify 100 support tickets | 0.15¢ | 0.42¢ | 2.7x |
| Review 500 lines of code | 0.75¢ | 0.83¢ | 1.1x |
| Write a SQL query | 0.10¢ | 0.13¢ | 1.3x |
| Run a 50-step coding agent | 8.7¢ | 9.2¢ | 1.1x |
Specs and list prices
| GPT-5.4 Nano | Qwen3.8 Flash | |
|---|---|---|
| Input per 1M | $0.20 | $0.15 |
| Output per 1M | $1.25 | $0.47 |
| Cached input per 1M | $0.02 | $0.016 |
| Context | 400K | 1M |
| Providers | 4 | 2 |
| Reasoning | default | none |
| LMArena Text | 1401 | - |
| Released | Mar 17, 2026 | Aug 26, 2026 |
FAQ
Is GPT-5.4 Nano or Qwen3.8 Flash cheaper?
GPT-5.4 Nano is cheaper on 10 of 14 tasks and about 1.4x cheaper on average (geometric mean).
How much does GPT-5.4 Nano cost compared to Qwen3.8 Flash per token?
GPT-5.4 Nano: $0.20 input and $1.25 output per 1M tokens. Qwen3.8 Flash: $0.15 input and $0.47 output. Token prices alone mislead - models use different numbers of tokens for the same task, which is what the task table shows.