autonomous long-horizon coding & professional work.
Qwen 3.8 Max beside its stablemates and nearest rivals. The ◆ marks the best value in each column across every row shown.
| Model | Intelligence | Coding | Speed | Input $/M | Output $/M | Blended $/M | Context |
|---|---|---|---|---|---|---|---|
| Qwen 3.7 Max | 46◆ | 66◆ | 199.8◆ | $2.50◆ | $7.50◆ | $3.75◆ | 1M◆ |
| Qwen 3.6 Max Preview | 40◆ | —◆ | —◆ | $1.30◆ | $7.80◆ | $2.92◆ | 200K◆ |
| Qwen 3.6 Plus | 39.6◆ | 54.5◆ | 55.3◆ | $0.50◆ | $3◆ | $1.13◆ | 200K◆ |
| Qwen 3.7 Plus | 39◆ | 55.9◆ | 52◆ | $0.40◆ | $1.60◆ | $0.70◆ | 1M◆ |
| Qwen 3.8 Max ◆ | —◆ | —◆ | —◆ | $2◆ | $6◆ | $3◆ | 1M◆ |
| Qwen 3.7 Flash | —◆ | —◆ | —◆ | $0.03◆ | $0.13◆ | $0.06◆ | 1M◆ |
| DeepSeek V4 Propinned | 44.3◆ | 59.4◆ | 59.3◆ | $0.43◆ | $0.87◆ | $0.54◆ | 1M◆ |
| DeepSeek V4 Flashpinned | 49.9◆ | 69.1◆ | —◆ | $0.14◆ | $0.28◆ | $0.18◆ | 1M◆ |
How far a month of credits goes on Qwen 3.8 Max.
Real coding tasks priced end to end on Qwen 3.8 Max, from a quick lookup to a full-repo agent run.
| Task | Tokens in · out | Qwen 3.8 Max | GPT-5.6 Terra | Claude Haiku 4.5 |
|---|---|---|---|---|
| Quick lookup / one-liner | 8K · 1K | $0.0094 | $0.0075 | $0.0065 |
| Review a 500-line PR | 60K · 4K | $0.07 | $0.05 | $0.04 |
| Fix a bug (agent loop) | 180K · 12K | $0.20 | $0.13 | $0.12 |
| Refactor a module | 320K · 20K | $0.36 | $0.23 | $0.21 |
| Full-repo agent run | 900K · 45K | $0.84 | $0.54 | $0.49 |
$2/M input and $6/M output, cache reads $0.25/M. In an agent loop most input is cache-read, so the effective input rate is about $0.78/M.cmd --model qwen/qwen3.8-max, or type /model in a session and pick it. You can switch mid-session without losing context.Command Code is the AI coding agent that continuously learns your taste. Start for $1.