- Size
- 31.3B parameterscounted from the safetensors header on the Hub
- Licence
- custom, read it, Apache License 2.0Our reading, check the licence: A custom licence we have not read for this model: other makers use the same name for different texts. Read the model's own licence file before you use it.Licence text
- On the Hub since
- 12 Mar 2026the date the repo was created
- Context length
- 256k tokensfrom the model's config
- Kind
- Base model (no base model declared on the Hub)Gemma4ForConditionalGeneration
- On Hugging Face
- google/gemma-4-31B44k downloads in the last 30 days, 608 likes
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: 1.4 GB, worked out from the model's config (google/gemma-4-31B), at 16-bit precision. Choose a machine above to see a verdict for each format. Method
GGUF for llama.cpp, Ollama, LM Studio and others
mradermacher/gemma-4-31B-i1-GGUF (converted by mradermacher), 18k downloads. Sizes are the real files on the Hub, all parts added up.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| IQ3_XXS | 11.2 GBfile size | 13.2 GB11.2 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:IQ3_XXSOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:IQ3_XXSDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-IQ3_XXS.gguf" --local-dir gemma-4-31B-i1-GGUF |
| IQ3_XS | 12.2 GBfile size | 14.1 GB12.2 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:IQ3_XSOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:IQ3_XSDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-IQ3_XS.gguf" --local-dir gemma-4-31B-i1-GGUF |
| IQ3_S | 12.8 GBfile size | 14.7 GB12.8 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:IQ3_SOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:IQ3_SDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-IQ3_S.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q3_K_S | 12.8 GBfile size | 14.7 GB12.8 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q3_K_SOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q3_K_SDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q3_K_S.gguf" --local-dir gemma-4-31B-i1-GGUF |
| IQ3_M | 13.4 GBfile size | 15.3 GB13.4 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:IQ3_MOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:IQ3_MDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-IQ3_M.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q3_K_M | 14.2 GBfile size | 16.1 GB14.2 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q3_K_MOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q3_K_MDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q3_K_M.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q3_K_L | 15.5 GBfile size | 17.4 GB15.5 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q3_K_LOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q3_K_LDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q3_K_L.gguf" --local-dir gemma-4-31B-i1-GGUF |
| IQ4_XS | 15.6 GBfile size | 17.5 GB15.6 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:IQ4_XSOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:IQ4_XSDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-IQ4_XS.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q4_0 | 16.5 GBfile size | 18.4 GB16.5 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q4_0Ollama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q4_0Download hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q4_0.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q4_K_S | 16.5 GBfile size | 18.4 GB16.5 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q4_K_SOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q4_K_SDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q4_K_S.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q4_K_M | 17.4 GBfile size | 19.3 GB17.4 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q4_K_MOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q4_K_MDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q4_K_M.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q4_1 | 18.1 GBfile size | 20.0 GB18.1 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q4_1Ollama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q4_1Download hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q4_1.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q5_K_S | 19.8 GBfile size | 21.8 GB19.8 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q5_K_SOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q5_K_SDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q5_K_S.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q5_K_M | 20.3 GBfile size | 22.3 GB20.3 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q5_K_MOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q5_K_MDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q5_K_M.gguf" --local-dir gemma-4-31B-i1-GGUF |
| Q6_K | 23.5 GBfile size | 25.4 GB23.5 + 1.4 + 0.50 | Commandsllama.cpp llama-server -hf mradermacher/gemma-4-31B-i1-GGUF:Q6_KOllama ollama run hf.co/mradermacher/gemma-4-31B-i1-GGUF:Q6_KDownload hf download mradermacher/gemma-4-31B-i1-GGUF --include "gemma-4-31B.i1-Q6_K.gguf" --local-dir gemma-4-31B-i1-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLX 2bitmajentik/gemma-4-31B-TurboQuant-MLX-2bit | 10.0 GBfile size | 11.9 GB10.0 + 1.4 + 0.50 | Download page |
| MLX 4bitmajentik/gemma-4-31B-TurboQuant-MLX-4bit | 17.1 GBfile size | 19.1 GB17.1 + 1.4 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| BF16original weights | 58.3 GBfile size | 60.2 GB58.3 + 1.4 + 0.50 |
hf download google/gemma-4-31B --local-dir gemma-4-31BOther conversions (14; sizes not read)
- axteam/gemma-4-31B-AWQ-8bit AWQ, 8bit, 336 downloads
- axteam/gemma-4-31B-FP8-dynamic FP8, 329 downloads
- axteam/gemma-4-31B-NVFP4 NVFP4, 323 downloads
- axteam/gemma-4-31B-NVFP4A16 NVFP4, 290 downloads
- axteam/gemma-4-31B-AWQ-4bit AWQ, 4bit, 248 downloads
- intel-ai/gemma-4-31B-MXFP8-FP8KV-CT-AutoRound Other, 96 downloads
- intel-ai/gemma-4-31B-MXFP8-FP8KV-FP8Attn-CT-AutoRound Other, 93 downloads
- lokeshe09/gemma-4-31B-it-INT4-W4A16 Other, int4, 77 downloads
- Xylolabs/gemma-4-31B-Tenstorrent-Tensix-BFP4-MLP Other, 61 downloads
- lokeshe09/gemma-4-31B-it-FP8-Dynamic FP8, 60 downloads
- Smartlearners/gemma-4-31B-it-FP8-Dynamic FP8, 47 downloads
- Spinf/gemma-4-31b-fp8 FP8, 42 downloads
- livadies/gemma-4-31B-Base-Ghetto-NF4 bitsandbytes, 28 downloads
- livadies/gemma-4-31B-Ghetto-NF4 bitsandbytes, 16 downloads
The Hub tags 46 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
Repos that declare this model as their base, most downloaded first. Listing is not a recommendation; unreviewed ones have not been checked.
- google/gemma-4-31B-it31.3B6.1MReviewed
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, 13:43 UTC. The repo. Found a mistake on this page? Tell us.