Gemma 4 12B
Gemma 4 12BRankingin thinking mode?
Alternatives?
CheaperSolar Mini 4+64%-42%$0.37No image input
BetterGPT-6 Lunalow+29%+33%$0.005Function calling only with none · Chat Completions+3
OpenQwen3.8 27Bxhigh×2.2+537%$1.01Memory (4-bit): 6 GB → 17 GB+5
CheaperLing 3.0 FlashthinkingOpen+51%-40%$0.06 / $0.18Memory (4-bit): 6 GB → 76 GB+1CheaperLing 3.0 Flash FinthinkingOpen+44%-40%$0.06 / $0.18Memory (4-bit): 6 GB → 76 GB+1Solar Mini 4+64%-42%$0.37No image inputLing 3.0 FlashthinkingOpen+51%-40%$0.06 / $0.18Memory (4-bit): 6 GB → 76 GB+1Ling 3.0 Flash FinthinkingOpen+44%-40%$0.06 / $0.18Memory (4-bit): 6 GB → 76 GB+1Ling 3.0 Flash VLthinkingOpen+40%-40%$0.06 / $0.18Memory (4-bit): 6 GB → 75 GBGPT OSS 120BhighOpen+25%-60%$0.092Context: 256K → 131K+2GPT-6 Lunalow+29%+33%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow+27%+200%$0.01
MiMo v2.5thinkingOpen+74%+17%$0.019Memory (4-bit): 6 GB → 186 GB
Claude Haiku 5.5low+72%+33%$0.024Remove sampling parameters+5GPT-5 Minihigh+37%+358%$0.053Qwen3.8 27BxhighOpen×2.2+537%$1.01Memory (4-bit): 6 GB → 17 GB+5Qwen 3.5 27BthinkingOpen+53%+450%$0.30 / $2.40Memory (4-bit): 6 GB → 17 GBQwen3.6 27BthinkingOpen+47%+800%$0.62Memory (4-bit): 6 GB → 17 GB
Muse Glimmer 30BhighOpen+40%+233%$0.058Context: 256K → 131K+1Gemma 4 26B A4BthinkingOpen+23%+75%$0.15 / $0.60Memory (4-bit): 6 GB → 15 GBGemini 4 Argonhigh×3.6—*$1.99Synchronous output stops at 65,536+2Gemini 3.1 Pro Previewthinking×3—$1.30Cannot use both thinking controlsGemini 3.8 Flashmedium×2.7—*$0.93Sampling parameters are ignored+1Gemini 3.6 Flashhigh×2.6—*$1.60Gemma 4 31BthinkingOpen+51%—$0.10 / $0.34Memory (4-bit): 6 GB → 19 GBGemini 2.5 Prothinking+44%—$0.33Gemini 2.5 Pro Preview (May' 25)thinking+27%——No image inputGemma 4 26B A4BthinkingOpen+23%—$0.15 / $0.60Memory (4-bit): 6 GB → 15 GBAll Google models →Run it yourself
estimateBenchmarks
all categoriesFAQ
How much does Gemma 4 12B cost?›
$0.10 per 1M input tokens and $0.30 per 1M output tokens (cheapest host: SiliconFlow).
Is Gemma 4 12B free to use?›
There is no free API; the cheapest paid price is $0.10 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Gemma 4 12B, and what is it best at?›
#143 of 334 in Chat & Knowledge, #95 of 239 in Coding, #122 of 175 in Agents & Tool Use, #67 of 95 in Writing & Language, #71 of 124 in Vision & Documents. Strongest at: Instructions.
Is Gemma 4 12B better than R1 0528?›
On Humanity's Last Exam, Gemma 4 12B scores 15.7 vs 15.8 — about the same.
How can I use Gemma 4 12B?›
Through the API of Featherless, SiliconFlow.
What is the context window of Gemma 4 12B?›
256K tokens.
When was Gemma 4 12B released? Is there a newer version?›
Released 2026-06-03. It is the current version in its line.
Is Gemma 4 12B open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Gemma 4 12B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Gemma 4 12B read images?›
It accepts text and images, audio as input.
What is Gemma 4 12B?›
Gemma 4 12B is an open-weights language model from Google, released Jun 2026.
What are the best alternatives to Gemma 4 12B?›
Solar Mini 4 (cheaper), GPT-6 Luna (better), Qwen3.8 27B (open weights).