| Claude Haiku 4.5 | gpt-oss-20b | |
|---|---|---|
| Input, per 1M tokens | $1.00 | $0.018 |
| Output, per 1M tokens | $5.00 | $0.09 |
| Context window | 200K | 131K |
| Classification or routing (600 tokens in, 10 out), per 1,000 | $0.65 | $0.0117 |
| Chat reply (2,000 in, 400 out), per 1,000 | $4.00 | $0.072 |
| Retrieval-augmented answer (8,000 in, 500 out), per 1,000 | $10.50 | $0.189 |
| Confidence signal | no | yes |
| Tool calling | yes | yes |
| BANKING77, correct of 400 | 301 | 313 |
Which one should carry your traffic?
If gpt-oss-20b answers your requests as well as Claude Haiku 4.5, moving that traffic saves 98% 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 Claude Haiku 4.5.
Claude Haiku 4.5 profile · gpt-oss-20b profile · Measure it on your traffic