- Size
- 4.3B parameterscounted from the safetensors header on the Hub
- Licence
- custom, read it, Health AI Developer Foundations Terms of UseOur 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
- 7 Jan 2026the date the repo was created
- Context length
- Not known
- Kind
- Base model (no base model declared on the Hub)Gemma3ForConditionalGeneration
- On Hugging Face
- google/medgemma-1.5-4b-it210k downloads in the last 30 days, 1.2k 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: 0.27 GB, worked out from the model's config (MedGemmaImpact/medgemma-1.5-4b-it-fp8), 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
unsloth/medgemma-1.5-4b-it-GGUF (converted by unsloth), 20k downloads. Sizes are the real files on the Hub, all parts added up. For image input you also need the projector file mmproj-BF16.gguf (0.79 GB), not counted below.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| UD-IQ3_XXS | 1.6 GBfile size | 2.4 GB1.6 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:UD-IQ3_XXSOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:UD-IQ3_XXSDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-UD-IQ3_XXS.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q3_K_S | 1.8 GBfile size | 2.6 GB1.8 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q3_K_SOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q3_K_SDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q3_K_S.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q3_K_M | 2.0 GBfile size | 2.7 GB2.0 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q3_K_MOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q3_K_MDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q3_K_M.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| UD-Q3_K_XL | 2.0 GBfile size | 2.8 GB2.0 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:UD-Q3_K_XLOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:UD-Q3_K_XLDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-UD-Q3_K_XL.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| IQ4_XS | 2.1 GBfile size | 2.9 GB2.1 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:IQ4_XSOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:IQ4_XSDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-IQ4_XS.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| IQ4_NL | 2.2 GBfile size | 3.0 GB2.2 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:IQ4_NLOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:IQ4_NLDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-IQ4_NL.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q4_0 | 2.2 GBfile size | 3.0 GB2.2 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q4_0Ollama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q4_0Download hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q4_0.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q4_K_S | 2.2 GBfile size | 3.0 GB2.2 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q4_K_SOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q4_K_SDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q4_K_S.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q4_K_M | 2.3 GBfile size | 3.1 GB2.3 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q4_K_MOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q4_K_MDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q4_K_M.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| UD-Q4_K_XL | 2.4 GBfile size | 3.1 GB2.4 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:UD-Q4_K_XLOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:UD-Q4_K_XLDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-UD-Q4_K_XL.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q4_1 | 2.4 GBfile size | 3.2 GB2.4 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q4_1Ollama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q4_1Download hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q4_1.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q5_K_S | 2.6 GBfile size | 3.3 GB2.6 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q5_K_SOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q5_K_SDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q5_K_S.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q5_K_M | 2.6 GBfile size | 3.4 GB2.6 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q5_K_MOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q5_K_MDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q5_K_M.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| UD-Q5_K_XL | 2.6 GBfile size | 3.4 GB2.6 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:UD-Q5_K_XLOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:UD-Q5_K_XLDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-UD-Q5_K_XL.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q6_K | 3.0 GBfile size | 3.7 GB3.0 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q6_KOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q6_KDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q6_K.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| UD-Q6_K_XL | 3.3 GBfile size | 4.1 GB3.3 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:UD-Q6_K_XLOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:UD-Q6_K_XLDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-UD-Q6_K_XL.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| Q8_0 | 3.8 GBfile size | 4.6 GB3.8 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:Q8_0Ollama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:Q8_0Download hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-Q8_0.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| UD-Q8_K_XL | 4.8 GBfile size | 5.6 GB4.8 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:UD-Q8_K_XLOllama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:UD-Q8_K_XLDownload hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-UD-Q8_K_XL.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
| BF16 | 7.2 GBfile size | 8.0 GB7.2 + 0.27 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/medgemma-1.5-4b-it-GGUF:BF16Ollama ollama run hf.co/unsloth/medgemma-1.5-4b-it-GGUF:BF16Download hf download unsloth/medgemma-1.5-4b-it-GGUF --include "medgemma-1.5-4b-it-BF16.gguf" --local-dir medgemma-1.5-4b-it-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLXponakilan/medgemma-1.5-4b-it-as | 2.0 GBfile size | 2.8 GB2.0 + 0.27 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| BF16original weights | 8.0 GBfile size | 8.8 GB8.0 + 0.27 + 0.50 |
hf download google/medgemma-1.5-4b-it --local-dir medgemma-1.5-4b-itOther conversions (13; sizes not read)
- MedGemmaImpact/medgemma-1.5-4b-it-fp8 FP8, 1.9k downloads
- unsloth/medgemma-1.5-4b-it-bnb-4bit bitsandbytes, 4bit, 769 downloads
- unsloth/medgemma-1.5-4b-it-unsloth-bnb-4bit bitsandbytes, 4bit, 739 downloads
- kamalkraj/medgemma-1.5-4b-it-executorch Other, 93 downloads
- suhaibrashid17/medgemma-1.5-4b-it-ONNX-webgpu Other, 65 downloads
- MedGemmaImpact/medgemma-1.5-4b-it-nvfp4 NVFP4, 37 downloads
- geeek/medgemma-1.5-4b-it-ONNX Other, 35 downloads
- ivangalkindeveloper/medgemma-1.5-4b-it-nvfp4 NVFP4, 34 downloads
- MedGemmaImpact/medgemma-1.5-4b-it-awq AWQ, 31 downloads
- Ashley101179/medgemma-1.5-4b-it-4bit-autoround GPTQ, 4bit, 21 downloads
- EmogheneAkpomedaye/MedGEMMA-1.5-4B-IT-ExecuTorch Other, 21 downloads
- CraneAILabs/medgemma-1.5-4b-it-nf4-blackwell bitsandbytes, 11 downloads
- aab20abdullah/medgemma-1.5-4b-it-4bit-nf4 bitsandbytes, 4bit, 10 downloads
The Hub tags 49 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.
- unsloth/medgemma-1.5-4b-it4.3B4.5kUnreviewed
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.