Qwen3.8 27B is cheaper than GPT-5.4 Pro at $0.4/M vs $30/M input tokens.
| Spec | GPT-5.4 Pro | Qwen3.8 27B |
|---|---|---|
| Input price | $30/M tokens | $0.4/M tokens |
| Output price | $180/M tokens | $3/M tokens |
| Context window | 1.1M tokens | 262K tokens |
| Weights | Closed | Open |
| Free API (OpenRouter) | No | No |
| Released | Mar 2026 | Aug 2026 |
Not enough votes to call it. On the specs, GPT-5.4 Pro has the edge: bigger model tier, bigger context window, major provider backing.
Qwen3.8 27B costs 60x less per token.
Reviewing agent-written code?See a Brief PR report
Qwen3.8 27B is cheaper on both: 75× input, 60× output.
Per million tokens. Prices and uptime via OpenRouter, checked 13 Sep 2026.
GPT-5.4 Pro uses 2.1x more headings