Qwen3.8 27B is cheaper than GPT-5.3-Codex at $0.4/M vs $1.75/M input tokens.
| Spec | GPT-5.3-Codex | Qwen3.8 27B |
|---|---|---|
| Input price | $1.75/M tokens | $0.4/M tokens |
| Output price | $14/M tokens | $3/M tokens |
| Context window | 400K tokens | 262K tokens |
| Released | Feb 2026 | Aug 2026 |
Not enough votes to call it. On the specs, GPT-5.3-Codex has the edge: bigger model tier, bigger context window, major provider backing.
Qwen3.8 27B costs 4.7x less per token.
Qwen3.8 27B is cheaper on both — 4.4× input, 4.7× output
GPT-5.3-Codex uses 15.8x more emoji
Qwen3.8 27B is cheaper than GPT-5.3-Codex at $0.4/M vs $1.75/M input tokens.
| Spec | GPT-5.3-Codex | Qwen3.8 27B |
|---|---|---|
| Input price | $1.75/M tokens | $0.4/M tokens |
| Output price | $14/M tokens | $3/M tokens |
| Context window | 400K tokens | 262K tokens |
| Released | Feb 2026 | Aug 2026 |
Not enough votes to call it. On the specs, GPT-5.3-Codex has the edge: bigger model tier, bigger context window, major provider backing.
Qwen3.8 27B costs 4.7x less per token.
Qwen3.8 27B is cheaper on both — 4.4× input, 4.7× output
GPT-5.3-Codex uses 15.8x more emoji