gemma-3-4b-it vs Qwen3.8-27B-iMatrix-NVFP4-MTP-GGUF

Specs, VRAM requirements and download trends — updated 10 November 2026.

google · A · Open weights

Params

4.3 B

Context

—

30d

1.0 M

gemma licence · conditional

cdiamond · B · Open weights

Params

—

Context

—

30d

3.3 M

apache-2.0 · commercial OK

Specification comparison

Differences are highlighted; identical values are muted.

Specification	gemma-3-4b-it	Qwen3.8-27B-iMatrix-NVFP4-MTP-GGUF
Deployment	Open weights · also via API	Open weights
Parameters	4.3 B	—
Architecture	Gemma3ForConditionalGeneration	—
Context length	131,072	—
Licence	gemma	apache-2.0
Commercial use	Conditional	Allowed
Languages	—	—
Downloads 30d	1,008,800	3,260,065
Downloads all time	28.1 M	9.6 M
Quantizations on the Hub	SAFETENSORS	GGUF
Gated	Yes — accept licence	No
First seen on the Hub	2025-02-20	2026-08-17
API input price / 1M tokens	$0.05	not on OpenRouter
API output price / 1M tokens	$0.10	not on OpenRouter
Est. cost per 1M input + 1M output tokens	$0.15	not on OpenRouter
Cached input / 1M tokens	—	not on OpenRouter
AA Intelligence Index	—	not on OpenRouter
AA Coding Index	2.7	not on OpenRouter
AA Agentic Index	—	not on OpenRouter
Max output tokens	16,384	not on OpenRouter
API providers	2	not on OpenRouter
Tool calling	No	not on OpenRouter
Reasoning mode	No	not on OpenRouter

Estimated cost of 1M input + 1M output tokens via OpenRouter

gemma-3-4b-it $0.15

Daily snapshots, last 90 days (13 Aug – 10 Nov)

5.0 M950 K

gemma-3-4b-it Qwen3.8-27B-iMatrix-NVFP4-MTP-GGUF

On RTX 4090 · 24 GB · 8K context unless noted

gemma-3-4b-it · fp16 10.6 GB / 24 GB

✅ Runs comfortably