| DeepSeek V4 Pro 0423 | Gemini 3.5 Flash Lite | |
|---|---|---|
| Input, per 1M tokens | $0.955 | $0.30 |
| Output, per 1M tokens | $1.91 | $2.50 |
| Context window | 1.05M | 1.05M |
| Classification or routing (600 tokens in, 10 out), per 1,000 | $0.592 | $0.205 |
| Chat reply (2,000 in, 400 out), per 1,000 | $2.67 | $1.60 |
| Retrieval-augmented answer (8,000 in, 500 out), per 1,000 | $8.60 | $3.65 |
| Confidence signal | yes | no |
| Tool calling | yes | yes |
| BANKING77, correct of 400 | not measured | 305 |
Which one should carry your traffic?
If Gemini 3.5 Flash Lite answers your requests as well as DeepSeek V4 Pro 0423, moving that traffic saves 40% on chat-shaped requests. Averages cannot tell you whether it does for your requests. Gemini 3.5 Flash Lite does not expose a confidence signal through OpenRouter, so Gimbix would test it as a full replacement rather than a first answer.
DeepSeek V4 Pro 0423 profile · Gemini 3.5 Flash Lite profile · Measure it on your traffic