| gpt-oss-20b | GLM 5.3 Flash | |
|---|---|---|
| Input, per 1M tokens | $0.018 | $0.15 |
| Output, per 1M tokens | $0.09 | $0.50 |
| Context window | 131K | 1.31M |
| Classification or routing (600 tokens in, 10 out), per 1,000 | $0.0117 | $0.095 |
| Chat reply (2,000 in, 400 out), per 1,000 | $0.072 | $0.50 |
| Retrieval-augmented answer (8,000 in, 500 out), per 1,000 | $0.189 | $1.45 |
| Confidence signal | yes | yes |
| Tool calling | yes | yes |
| BANKING77, correct of 400 | 313 | 320 |
Which one should carry your traffic?
If gpt-oss-20b answers your requests as well as GLM 5.3 Flash, moving that traffic saves 86% on chat-shaped requests. Averages cannot tell you whether it does for your requests. gpt-oss-20b exposes a confidence signal, so Gimbix can let it answer first and hand unsure cases to GLM 5.3 Flash.
gpt-oss-20b profile · GLM 5.3 Flash profile · Measure it on your traffic