| Claude Haiku 4.5 | Grok 4.7 | |
|---|---|---|
| Input, per 1M tokens | $1.00 | $1.60 |
| Output, per 1M tokens | $5.00 | $4.80 |
| Context window | 200K | 500K |
| Classification or routing (600 tokens in, 10 out), per 1,000 | $0.65 | $1.01 |
| Chat reply (2,000 in, 400 out), per 1,000 | $4.00 | $5.12 |
| Retrieval-augmented answer (8,000 in, 500 out), per 1,000 | $10.50 | $15.20 |
| Confidence signal | no | yes |
| Tool calling | yes | yes |
| BANKING77, correct of 400 | 301 | not measured |
Which one should carry your traffic?
If Claude Haiku 4.5 answers your requests as well as Grok 4.7, moving that traffic saves 22% on chat-shaped requests. Averages cannot tell you whether it does for your requests. Claude Haiku 4.5 does not expose a confidence signal through OpenRouter, so Gimbix would test it as a full replacement rather than a first answer.
Claude Haiku 4.5 profile · Grok 4.7 profile · Measure it on your traffic