Llama-3.2-1B-Instruct vs Qwen3-4B-Instruct-2507-NVFP4
Specs, VRAM requirements and download trends — updated 8 August 2026.
meta-llama · A · Open weights
Params
1.2 B
Context
—
30d
1.3 M
llama3.2 licence · conditional
llmat · B · Open weights
Params
2.8 B
Context
262,144
30d
161 K
apache-2.0 · commercial OK
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | Llama-3.2-1B-Instruct | Qwen3-4B-Instruct-2507-NVFP4 |
|---|---|---|
| Deployment | Open weights · also via API | Open weights |
| Parameters | 1.2 B | 2.8 B |
| Architecture | LlamaForCausalLM | Qwen3ForCausalLM |
| Context length | 60,000 | 262,144 |
| Licence | llama3.2 | apache-2.0 |
| Commercial use | Conditional | Allowed |
| Languages | 8 | 1 |
| Downloads 30d | 1,331,481 | 160,914 |
| Downloads all time | 109.8 M | 269 K |
| Quantizations on the Hub | SAFETENSORS | SAFETENSORS |
| Gated | Yes — accept licence | No |
| First seen on the Hub | 2024-09-18 | 2025-08-27 |
| API input price / 1M tokens | $0.027 | not on OpenRouter |
| API output price / 1M tokens | $0.20 | not on OpenRouter |
| Est. cost per 1M input + 1M output tokens | $0.23 | not on OpenRouter |
| Cached input / 1M tokens | — | not on OpenRouter |
| AA Intelligence Index | — | not on OpenRouter |
| AA Coding Index | — | not on OpenRouter |
| AA Agentic Index | — | not on OpenRouter |
| Max output tokens | 54,000 | not on OpenRouter |
| API providers | 1 | not on OpenRouter |
| Tool calling | No | not on OpenRouter |
| Reasoning mode | No | not on OpenRouter |
API cost
Estimated cost of 1M input + 1M output tokens via OpenRouter
Llama-3.2-1B-Instruct
$0.23
Download trend
Daily snapshots, last 90 days (13 Aug – 10 Nov)
9.3 M161 K
Llama-3.2-1B-Instruct
Qwen3-4B-Instruct-2507-NVFP4
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
Llama-3.2-1B-Instruct · fp16
3.4 GB / 24 GB
Qwen3-4B-Instruct-2507-NVFP4 · fp16 @ 256K ctx
18.0 GB / 24 GB