All models › google
Arena score1358#199 of 416 in Chat, paid models included
Size27.4 Bparameters
4-bit download17 GBweights only, Q4_K_M
Max context128Ktokens, from the model config
Where to use it for free
- The lab's own chat: aistudio.google.com (free with daily limits)
- Download and run it at home: official weights · GGUF files for Ollama and LM Studio
Arena scores
Rank among all models, paid ones included. Source: Arena leaderboard dataset.
Speed on popular hardware
| Hardware | GPU memory | Fits? | Speed |
|---|---|---|---|
| RTX 3060 12 GB | 12 GB | GPU + 64 GB RAM | 6.4 tok/s |
| RTX 4060 Ti 16 GB | 16 GB | GPU + 64 GB RAM | 7.3 tok/s |
| RTX 3090 | 24 GB | Fits in GPU | 33 tok/s |
| RTX 4090 | 24 GB | Fits in GPU | 35 tok/s |
| RTX 5090 | 32 GB | Fits in GPU | 63 tok/s |
| RX 7900 XTX | 24 GB | Fits in GPU | 34 tok/s |
| 2 × RTX 3090 | 48 GB | Fits in GPU | 32 tok/s |
| 4 × RTX 3090 | 96 GB | Fits in GPU | 30 tok/s |
| Mac M4 Max 128 GB | 96 GB | Fits in GPU | 19 tok/s |
| Ryzen AI Max+ 395 128 GB | 96 GB | Fits in GPU | 9.0 tok/s |
| 8 × H100 SXM | 640 GB | Fits in GPU | 97 tok/s |
Estimate at 4-bit with 32K context, 64 GB of DDR5 for offload. Speed ≈ memory bandwidth × 70% ÷ data read per token; real numbers vary by about ±30%.
GPU memory needed
| Level | 8K context | 32K context | 128K context | Quality vs full |
|---|---|---|---|---|
| Q3_K_M | 15 GB | 18 GB | 27 GB | +7.2% noticeable loss |
| Q4_K_M | 18 GB | 21 GB | 29 GB | +1.9% small loss |
| Q5_K_M | 21 GB | 23 GB | 32 GB | +0.6% very small loss |
| Q6_K | 24 GB | 26 GB | 35 GB | +0.2% near-lossless |
| Q8_0 | 31 GB | 33 GB | 42 GB | ≈ 0% same as full |
| F16 | 57 GB | 59 GB | 68 GB | ≈ 0% same as full |
Weights + context cache (from the model's architecture) + runtime. Quality: increase in perplexity against the full model, from llama.cpp tests, scaled by model size. Simulate your own setup.
Chat score over time
14001500
MayJunJulAugSeptOct
gemma-3-27b-it 1358Best paid model
Arena score from 17 Apr 2026 to 15 Oct 2026.
Vision score over time
11501200125013001350
MayJunJulAugSeptOct
gemma-3-27b-it 1164Best paid model
Arena score from 10 Apr 2026 to 9 Oct 2026.
Similar free models in Chat
- minimax-m2.5 1359
- Qwen3-Coder-480B-A35B-Instruct 1357
- intellect-3 1356
- granite-4.2-30b 1361
- ling-flash-2.0 1361
- glm-4.7-flash 1351