| Nova Lite 1.0 | Claude Haiku 4.5 | |
|---|---|---|
| Input, per 1M tokens | $0.06 | $1.00 |
| Output, per 1M tokens | $0.24 | $5.00 |
| Context window | 300K | 200K |
| Classification or routing (600 tokens in, 10 out), per 1,000 | $0.0384 | $0.65 |
| Chat reply (2,000 in, 400 out), per 1,000 | $0.216 | $4.00 |
| Retrieval-augmented answer (8,000 in, 500 out), per 1,000 | $0.60 | $10.50 |
| Confidence signal | no | no |
| Tool calling | yes | yes |
| BANKING77, correct of 400 | not measured | 301 |
Which one should carry your traffic?
If Nova Lite 1.0 answers your requests as well as Claude Haiku 4.5, moving that traffic saves 95% on chat-shaped requests. Averages cannot tell you whether it does for your requests. Nova Lite 1.0 does not expose a confidence signal through OpenRouter, so Gimbix would test it as a full replacement rather than a first answer.
Nova Lite 1.0 profile · Claude Haiku 4.5 profile · Measure it on your traffic