google · fine-tune of google/gemma-4-12B
gemma-4-12B-it
Runs on your machine
Choose a machine
We then show the largest format that fits, from real file sizes where we have them.
Scores
No independent scores for this model in Epoch AI's data. A score is only shown here when Epoch's name for a model matches this repo with certainty; an uncertain match is left out rather than guessed.
Claimed by the maker Claimed
The model card has no claimed scores in a machine-readable form (the card's model-index), so none are shown. The maker may give scores in the card text: read the card.
Other views, which we link to and do not copy: look this model up on Artificial Analysis and the Arena leaderboard.
Run it yourself
Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 0.44 GB, worked out from the model's config (google/gemma-4-12B-it), at 16-bit precision. Choose a machine above to see a verdict for each format. Method
weights context cache runtime
GGUF for llama.cpp, Ollama, LM Studio and others
unsloth/gemma-4-12b-it-GGUF (converted by unsloth), 2.6M downloads. Sizes are the real files on the Hub, all parts added up. Showing the common levels.
| Format | Size | Memory needed | Memory | Get it |
|---|---|---|---|---|
| Q4_K_M | 6.6 GBfile size | 7.6 GB6.6 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/gemma-4-12b-it-GGUF:Q4_K_MOllama ollama run hf.co/unsloth/gemma-4-12b-it-GGUF:Q4_K_MDownload hf download unsloth/gemma-4-12b-it-GGUF --include "gemma-4-12b-it-Q4_K_M.gguf" --local-dir gemma-4-12b-it-GGUF | |
| Q5_K_M | 7.8 GBfile size | 8.8 GB7.8 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/gemma-4-12b-it-GGUF:Q5_K_MOllama ollama run hf.co/unsloth/gemma-4-12b-it-GGUF:Q5_K_MDownload hf download unsloth/gemma-4-12b-it-GGUF --include "gemma-4-12b-it-Q5_K_M.gguf" --local-dir gemma-4-12b-it-GGUF | |
| Q6_K | 9.1 GBfile size | 10.1 GB9.1 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/gemma-4-12b-it-GGUF:Q6_KOllama ollama run hf.co/unsloth/gemma-4-12b-it-GGUF:Q6_KDownload hf download unsloth/gemma-4-12b-it-GGUF --include "gemma-4-12b-it-Q6_K.gguf" --local-dir gemma-4-12b-it-GGUF | |
| Q8_0 | 11.8 GBfile size | 12.7 GB11.8 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/gemma-4-12b-it-GGUF:Q8_0Ollama ollama run hf.co/unsloth/gemma-4-12b-it-GGUF:Q8_0Download hf download unsloth/gemma-4-12b-it-GGUF --include "gemma-4-12b-it-Q8_0.gguf" --local-dir gemma-4-12b-it-GGUF | |
| BF16 | 22.2 GBfile size | 23.1 GB22.2 + 0.44 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/gemma-4-12b-it-GGUF:BF16Ollama ollama run hf.co/unsloth/gemma-4-12b-it-GGUF:BF16Download hf download unsloth/gemma-4-12b-it-GGUF --include "gemma-4-12b-it-BF16.gguf" --local-dir gemma-4-12b-it-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Memory | Repo |
|---|---|---|---|---|
| MLX 4bitlmstudio-community/gemma-4-12B-it-MLX-4bit | 6.3 GBfile size | 7.2 GB6.3 + 0.44 + 0.50 | Download page | |
| MLX 8bitlmstudio-community/gemma-4-12B-it-MLX-8bit | 11.8 GBfile size | 12.8 GB11.8 + 0.44 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Memory | Repo |
|---|---|---|---|---|
| BF16original weights | 22.3 GBfile size | 23.2 GB22.3 + 0.44 + 0.50 |
hf download google/gemma-4-12B-it --local-dir gemma-4-12B-itOther conversions (12; sizes not read)
- google/gemma-4-12B-it-qat-q4_0-unquantized Other, by the maker, 9.0k downloads
- cyankiwi/gemma-4-12B-it-AWQ-INT4 AWQ, int4, 81k downloads
- unsloth/gemma-4-12b-it-NVFP4 NVFP4, 13k downloads
- aleada/Gemma-4-12B-it-W4A16 GPTQ, w4a16, 12k downloads
- RedHatAI/gemma-4-12B-it-FP8-Dynamic FP8, 5.2k downloads
- RedHatAI/gemma-4-12B-it-NVFP4 NVFP4, 4.0k downloads
- Intel/gemma-4-12B-it-int8-AutoRound Other, int8, 1.2k downloads
- coolthor/gemma-4-12B-it-FP8-dynamic FP8, 762 downloads
- axteam/gemma-4-12B-it-AWQ-4bit AWQ, 4bit, 454 downloads
- coolthor/gemma-4-12B-it-NVFP4A16 NVFP4, 445 downloads
- turboderp/gemma-4-12B-it-exl3 EXL3, 340 downloads
- axteam/gemma-4-12B-it-NVFP4 NVFP4, 293 downloads
The Hub tags 100 or more quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.
Run it by API
No provider on Hugging Face Inference Providers serves this model (as of 19 Nov 2026, 13:58 UTC). Other hosts may; we only list what the router reports.
Not built yet: one key for every open model, with BlockRun, and our own tested-quality score per provider. Until then, precision and measured speed are the best guide to a provider here.
Fine-tunes of this model
- AEON-7/Gemma-4-12B-it-AEON-Abliterated-K4-BF1612.0B70kUnreviewed
- unsloth/gemma-4-12b-it12.0B43kUnreviewed
Planned for this page Not built yet
- Try it in private chat
- A link to Solana AI's confidential chat for the models it serves. This model is not listed as served there.
- Fine-tune it
- Fine-tuning a few base models from this page and publishing the result here.
- Creators and earnings
- Pages for people who publish fine-tunes and adapters of this model, and what they earn.
- Fund and bounties
- Bounties for new open models, paid from a public fund. How the fund will work.
- Contribute compute
- How to give compute to open model work, with Akash.
Model data from the Hugging Face Hub 19 Nov 2026, 15:15 UTC. The repo. Found a mistake on this page? Tell us.