Qwen3.5-9B vs Nova 2 Lite
API pricing, benchmark indices and specs — updated 19 October 2026.
amazon · B · API-only
Input / 1M
$0.30
Output / 1M
$2.50
AA index
13.4
proprietary · 1M context
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | Qwen3.5-9B | Nova 2 Lite |
|---|---|---|
| Deployment | Open weights · also via API | API-only (closed weights) |
| Parameters | 9.7 B | n/a — API-only |
| Architecture | Qwen3_5ForConditionalGeneration | n/a — API-only |
| Context length | 262,144 | 1,000,000 |
| Licence | apache-2.0 | Proprietary |
| Commercial use | Allowed | Per API terms |
| Languages | — | n/a — API-only |
| Downloads 30d | 6,369,264 | n/a — API-only |
| Downloads all time | 71.5 M | n/a — API-only |
| Quantizations on the Hub | SAFETENSORS | n/a — API-only |
| Gated | No | n/a — API-only |
| First seen on the Hub | 2026-02-27 | n/a — API-only |
| API input price / 1M tokens | $0.17 | $0.30 |
| API output price / 1M tokens | $0.25 | $2.50 |
| Est. cost per 1M input + 1M output tokens | $0.42 | $2.80 |
| Cached input / 1M tokens | — | — |
| AA Intelligence Index | 13.3 | 13.4 |
| AA Coding Index | 23.5 | 23 |
| AA Agentic Index | — | — |
| Max output tokens | 235,929 | 65,535 |
| API providers | 5 | 1 |
| Tool calling | Yes | Yes |
| Reasoning mode | Yes | Yes |
API cost
Estimated cost of 1M input + 1M output tokens via OpenRouter
Qwen3.5-9B
$0.42
Nova 2 Lite
$2.80
Qwen3.5-9B is 6.7× cheaper at list price.
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
Qwen3.5-9B · fp16
23.2 GB / 24 GB