| Kimi K3 | Qwen3.8 Max (0902) | |
|---|---|---|
| Input, per 1M tokens | $3.00 | $2.00 |
| Output, per 1M tokens | $15.00 | $6.00 |
| Context window | 1.05M | 1M |
| Classification or routing (600 tokens in, 10 out), per 1,000 | $1.95 | $1.26 |
| Chat reply (2,000 in, 400 out), per 1,000 | $12.00 | $6.40 |
| Retrieval-augmented answer (8,000 in, 500 out), per 1,000 | $31.50 | $19.00 |
| Confidence signal | yes | yes |
| Tool calling | yes | yes |
Which one should carry your traffic?
If Qwen3.8 Max (0902) answers your requests as well as Kimi K3, moving that traffic saves 47% on chat-shaped requests. Averages cannot tell you whether it does for your requests. Qwen3.8 Max (0902) exposes a confidence signal, so Gimbix can let it answer first and hand unsure cases to Kimi K3.
Kimi K3 profile · Qwen3.8 Max (0902) profile · Measure it on your traffic