Qwen2.5-VL-3B-Instruct vs Kimi-K2.6
Specs, VRAM requirements and download trends — updated 28 July 2026.
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | Qwen2.5-VL-3B-Instruct | Kimi-K2.6 |
|---|---|---|
| Deployment | Open weights | Open weights · also via API |
| Parameters | 3.8 B | 1,058.6 B |
| Architecture | Qwen2_5_VLForConditionalGeneration | KimiK25ForConditionalGeneration |
| Context length | 128,000 | 262,144 |
| Licence | — | other |
| Commercial use | Unknown | Unknown |
| Languages | 1 | — |
| Downloads 30d | 1,736,195 | 834,406 |
| Downloads all time | 105.4 M | 7.7 M |
| Quantizations on the Hub | SAFETENSORS | SAFETENSORS |
| Gated | No | No |
| First seen on the Hub | 2025-01-26 | 2026-04-14 |
| API input price / 1M tokens | not on OpenRouter | $0.65 |
| API output price / 1M tokens | not on OpenRouter | $3.41 |
| Est. cost per 1M input + 1M output tokens | not on OpenRouter | $4.06 |
| Cached input / 1M tokens | not on OpenRouter | $0.15 |
| AA Intelligence Index | not on OpenRouter | — |
| AA Coding Index | not on OpenRouter | 61.8 |
| AA Agentic Index | not on OpenRouter | 20.5 |
| Max output tokens | not on OpenRouter | 235,929 |
| API providers | not on OpenRouter | 16 |
| Tool calling | not on OpenRouter | Yes |
| Reasoning mode | not on OpenRouter | Yes |
API cost
Estimated cost of 1M input + 1M output tokens via OpenRouter
Kimi-K2.6
$4.06
Download trend
Daily snapshots, last 90 days (13 Aug – 10 Nov)
7.2 M805 K
Qwen2.5-VL-3B-Instruct
Kimi-K2.6
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
Qwen2.5-VL-3B-Instruct · fp16 @ 125K ctx
17.5 GB / 24 GB
Kimi-K2.6 · fp16
814.0 GB / 24 GB