Gemma 4 E4B
Gemma 4 E4BRankingin thinking mode?
Alternatives?
CheaperGPT-6 Lunalow×2.4-45%$0.005Function calling only with none · Chat Completions+3IOpenK2 Horizon 7Bthinking×2.3+1,212%$0.30 / $1.20Memory (4-bit): 4 GB → 4.5 GB+1CheaperGPT OSS 20BlowOpen+12%-10%$0.018 / $0.09Memory (4-bit): 4 GB → 13 GB+1
CheaperGranite 4.2 3BthinkingOpensame-27%$0.006No image input
CheaperQwen3.5 0.8BthinkingOpen-31%-50%$0.01 / $0.05GPT-6 Lunalow×2.4-45%$0.005Function calling only with none · Chat Completions+3GPT OSS 20BlowOpen+12%-10%$0.018 / $0.09Memory (4-bit): 4 GB → 13 GB+1Granite 4.2 3BthinkingOpensame-27%$0.006No image inputQwen3.5 0.8BthinkingOpen-31%-50%$0.01 / $0.05GPT-6 Lunalow×2.4-45%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.4+20%$0.01
Mistral Small 4thinkingOpen+26%+84%$0.015Memory (4-bit): 4 GB → 72 GB
MiMo v2.5thinkingOpen×2.4+134%$0.019Memory (4-bit): 4 GB → 186 GBGranite 4.2 8BthinkingOpen+24%+189%$0.024Memory (4-bit): 4 GB → 4.4 GB+1Gemma 4 12BthinkingOpen+59%+275%$0.10 / $0.30Memory (4-bit): 4 GB → 6 GBSApriel v1.5 15B ThinkerOpen+55%——Memory (4-bit): 4 GB → 7.4 GBSApriel v1.6 15B ThinkerOpen+50%——Memory (4-bit): 4 GB → 7.4 GBGemini 4 Argonhigh×5.9+24,281%*$1.99Synchronous output stops at 65,536+2Gemini 3.8 Flashmedium×4.5+11,305%*$0.93Sampling parameters are ignored+1Gemini 3.6 Flashhigh×3.8+19,506%*$1.60Gemini 3.1 Pro Previewthinking×3.3+15,808%$1.30Cannot use both thinking controlsGemini 3.5 Flash Litehigh×2.5+2,261%$0.19Gemini 3 Flash Previewnon-thinking×2—$0.50 / $3Gemma 4 26B A4BthinkingOpen+87%—$0.15 / $0.60Memory (4-bit): 4 GB → 15 GBGemini 2.5 Prothinking+80%+3,946%$0.33All Google models →Run it yourself
estimateBenchmarks
all categoriesFAQ
How much does Gemma 4 E4B cost?›
$0.02 per 1M input tokens and $0.10 per 1M output tokens (cheapest host: DeepInfra). A typical benchmark task costs about $0.008 at default effort.
Is Gemma 4 E4B free to use?›
There is no free API; the cheapest paid price is $0.02 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Gemma 4 E4B, and what is it best at?›
#317 of 334 in Chat & Knowledge, #75 of 132 in Coding, #121 of 133 in Agents & Tool Use, #133 of 245 in Writing & Language, #109 of 124 in Vision & Documents.
Is Gemma 4 E4B better than Gemma 3n E4B Instruct?›
On AA Intelligence Index 4.3.2, Gemma 4 E4B scores 8.9 vs 4.8 — higher.
How can I use Gemma 4 E4B?›
Through the API of DeepInfra, Featherless.
What is the context window of Gemma 4 E4B?›
128K tokens.
When was Gemma 4 E4B released? Is there a newer version?›
Released 2026-04-03. It is the current version in its line.
Is Gemma 4 E4B open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Gemma 4 E4B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Gemma 4 E4B read images?›
It accepts text and images, audio as input.
What is Gemma 4 E4B?›
Gemma 4 E4B is an open-weights language model from Google, released Apr 2026, the successor to Gemma 3n E4B Instruct.
What are the best alternatives to Gemma 4 E4B?›
GPT-6 Luna (cheaper), K2 Horizon 7B (open weights).