Skip to content
Weights Not Walls
Menu

google

gemma-2-27b-it

Reviewedgated on the Hub
Size
27.2B parameterscounted from the safetensors header on the Hub
Licence
unknown, Gemma Terms of UseOur reading, check the licence: Commercial use is allowed with no user or revenue cap, but the Gemma Prohibited Use Policy (a separate document, not in this text) is part of the terms, and you must pass the use restrictions on as enforceable terms to anyone you distribute to. Models distilled from Gemma or trained on its outputs to perform like it count as Model Derivatives and stay under these terms, and Google may restrict use remotely if it believes the terms are broken. The page says these terms cover the models in its appendix (Gemma 1 to 3, 3n and variants) and that Gemma 4 has a separate licence.
On the Hub since
24 Jun 2024the date the repo was created
Context length
8k tokensas an API provider reports it
Fine-tune
of google/gemma-2-27bGemma2ForCausalLM
On Hugging Face
google/gemma-2-27b-it23k downloads in the last 30 days, 579 likes

Scores

5 independent results on 5 eval sets
GPQA diamondRe-run by Epoch AI
  • default setting36.5%±2.427 Jan 2025, log

Epoch's name: gemma-2-27b-it. Best closed model on this set: gpt-6.2-astra_max, 96.8% (OpenAI). All 341 results

MATH level 5Re-run by Epoch AI
  • default setting27.9%±1.027 Jan 2025, log

Epoch's name: gemma-2-27b-it. Best closed model on this set: gpt-5-2025-08-07_high, 98.1% (OpenAI). All 108 results

OTIS Mock AIME 2024-2025Re-run by Epoch AI
  • default setting1.4%±0.97 Mar 2025, log

Epoch's name: gemma-2-27b-it. Best closed model on this set: claude-fable-5_high, 100.0% (Anthropic). All 318 results

LiveBenchReported

Epoch's name: gemma-2-27b-it. Best closed model on this set: gemini-2.5-pro-exp-03-25, 82.3% (Google DeepMind). All 54 results

MMLUReported

Epoch's name: gemma-2-27b-it. Best closed model on this set: gpt-4o-2024-11-20, 88.1% (OpenAI). All 217 results

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.

Independent scores from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Each row links its own source. What the badges mean.

Run it yourself

file sizes read from the Hub 19 Nov 2026

Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 2.9 GB, worked out from the model's config (mbley/google-gemma-2-27b-it-AWQ), 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

bartowski/gemma-2-27b-it-GGUF (converted by bartowski), 16k downloads. Sizes are the real files on the Hub, all parts added up.

FormatSizeMemory neededGet it
IQ3_XXS10.0 GBfile size13.4 GB10.0 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:IQ3_XXS
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:IQ3_XXS
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-IQ3_XXS.gguf" --local-dir gemma-2-27b-it-GGUF
IQ3_XS10.8 GBfile size14.1 GB10.8 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:IQ3_XS
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:IQ3_XS
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-IQ3_XS.gguf" --local-dir gemma-2-27b-it-GGUF
Q3_K_S11.3 GBfile size14.7 GB11.3 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q3_K_S
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q3_K_S
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q3_K_S.gguf" --local-dir gemma-2-27b-it-GGUF
IQ3_M11.6 GBfile size15.0 GB11.6 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:IQ3_M
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:IQ3_M
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-IQ3_M.gguf" --local-dir gemma-2-27b-it-GGUF
Q3_K_M12.5 GBfile size15.9 GB12.5 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q3_K_M
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q3_K_M
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q3_K_M.gguf" --local-dir gemma-2-27b-it-GGUF
Q3_K_L13.5 GBfile size16.9 GB13.5 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q3_K_L
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q3_K_L
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q3_K_L.gguf" --local-dir gemma-2-27b-it-GGUF
Q3_K_XL13.8 GBfile size17.2 GB13.8 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q3_K_XL
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q3_K_XL
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q3_K_XL.gguf" --local-dir gemma-2-27b-it-GGUF
IQ4_XS13.8 GBfile size17.2 GB13.8 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:IQ4_XS
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:IQ4_XS
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-IQ4_XS.gguf" --local-dir gemma-2-27b-it-GGUF
Q4_K_S14.7 GBfile size18.0 GB14.7 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q4_K_S
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q4_K_S
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q4_K_S.gguf" --local-dir gemma-2-27b-it-GGUF
Q4_K_M15.5 GBfile size18.9 GB15.5 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q4_K_M
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q4_K_M
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q4_K_M.gguf" --local-dir gemma-2-27b-it-GGUF
Q4_K_L15.8 GBfile size19.1 GB15.8 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q4_K_L
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q4_K_L
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q4_K_L.gguf" --local-dir gemma-2-27b-it-GGUF
Q5_K_S17.6 GBfile size21.0 GB17.6 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q5_K_S
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q5_K_S
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q5_K_S.gguf" --local-dir gemma-2-27b-it-GGUF
Q5_K_M18.1 GBfile size21.5 GB18.1 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q5_K_M
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q5_K_M
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q5_K_M.gguf" --local-dir gemma-2-27b-it-GGUF
Q5_K_L18.3 GBfile size21.7 GB18.3 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q5_K_L
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q5_K_L
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q5_K_L.gguf" --local-dir gemma-2-27b-it-GGUF
Q6_K20.8 GBfile size24.2 GB20.8 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q6_K
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q6_K
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q6_K.gguf" --local-dir gemma-2-27b-it-GGUF
Q6_K_L21.1 GBfile size24.5 GB21.1 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q6_K_L
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q6_K_L
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q6_K_L.gguf" --local-dir gemma-2-27b-it-GGUF
Q8_027.0 GBfile size30.3 GB27.0 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q8_0
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q8_0
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q8_0.gguf" --local-dir gemma-2-27b-it-GGUF
Q8_0_L28.0 GBfile size31.4 GB28.0 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:Q8_0_L
Ollama
ollama run hf.co/bartowski/gemma-2-27b-it-GGUF:Q8_0_L
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-Q8_0_L.gguf" --local-dir gemma-2-27b-it-GGUF
F32101.4 GBfile size, 3 parts104.8 GB101.4 + 2.9 + 0.50
Commands
llama.cpp
llama-server -hf bartowski/gemma-2-27b-it-GGUF:F32
Download
hf download bartowski/gemma-2-27b-it-GGUF --include "gemma-2-27b-it-f32.gguf/gemma-2-27b-it-f32-*-of-00003.gguf" --local-dir gemma-2-27b-it-GGUF

MLX for Apple silicon

FormatSizeMemory neededRepo
MLX 4bitmlx-community/gemma-2-27b-it-bf16-4bit14.3 GBfile size17.6 GB14.3 + 2.9 + 0.50Download page
MLX 8bitmlx-community/gemma-2-27b-it-bf16-8bit26.9 GBfile size30.3 GB26.9 + 2.9 + 0.50Download page

Original weights size from the safetensors header, for vLLM, SGLang, Transformers

FormatSizeMemory neededRepo
BF16original weights50.7 GBfile size54.1 GB50.7 + 2.9 + 0.50
hf download google/gemma-2-27b-it --local-dir gemma-2-27b-it
Other conversions (8; sizes not read)

The Hub tags 60 quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.

Run it by API

Hugging Face Inference Providers, fetched 19 Nov 2026, 13:58 UTC

2 providers serve this model through the Hugging Face router. Prices are US dollars per million tokens. Precision is what the provider reports; a lower precision than the original weights can change answers. Live first, then cheapest output.

ProviderInOutContextPrecisionTokens/sFirst tokenUptime 24hToolsJSON
NextBit$0.65$0.658kint4430.6 s100.0%noyes
featherless-ai???not given???not givennot given

Through the router, with a Hugging Face token (the router bills your Hugging Face account). Change the name after the colon to pick another provider.

curl https://router.huggingface.co/v1/chat/completions \ -H "Authorization: Bearer $HF_TOKEN" -H "Content-Type: application/json" \ -d '{"model": "google/gemma-2-27b-it:nextbit", "messages": [{"role": "user", "content": "Hello"}]}'

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

from our snapshot of the Hub

Repos that declare this model as their base, most downloaded first. Listing is not a recommendation; unreviewed ones have not been checked.

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.