FitLocally by GPUwerk

Gemma · 2B

Can you run Gemma 2 2B Instruct locally?

Gemma 2 2B Instruct is a dense model with 2B parameters. At 4-bit or better it runs on 120 of the 120 hardware configurations listed here; the smallest is the RTX 5060 (8 GB). On a 128 GB DGX Spark it runs at Q4_K_M, at roughly 83 tokens per second (our est.).

Data as of 2 November 2026

Fastest machines for Gemma 2 2B Instruct

Every machine that runs it, fastest first. Filter by type; your machine is marked. Compare machines running it side by side

  1. 1RTX 5090Graphics card · 32 GB~373 tok/s ModelFit est.Q4_K_M · 2.8 of 31 GB
  2. 2RTX PRO 6000 BlackwellGraphics card · 96 GB~373 tok/s ModelFit est.Q4_K_M · 2.8 of 95 GB
  3. 32x RTX 5090Graphics card · 64 GB~336 tok/s ModelFit est.Q4_K_M · 2.8 of 62 GB
  4. 42x RTX PRO 6000 BlackwellGraphics card · 192 GB~336 tok/s ModelFit est.Q4_K_M · 2.8 of 190 GB
  5. 5RTX 4090Graphics card · 24 GB~263 tok/s ModelFit est.Q4_K_M · 2.8 of 23 GB
  6. 6RTX 5080Graphics card · 16 GB~262 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  7. 7RTX 5070 TiGraphics card · 16 GB~250 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  8. 82x RTX 4090Graphics card · 48 GB~244 tok/s ModelFit est.Q4_K_M · 2.8 of 46 GB
  9. 9M5 UltraMac · 96–512 GB · 5 variants~243 tok/s ModelFit est.Q4_K_M · 2.8 GB
  10. 10RTX 3090 TiGraphics card · 24 GB~241 tok/s ModelFit est.Q4_K_M · 2.8 of 23 GB
  11. 11RTX 3090Graphics card · 24 GB~230 tok/s ModelFit est.Q4_K_M · 2.8 of 23 GB
  12. 122x RTX 3090 TiGraphics card · 48 GB~225 tok/s ModelFit est.Q4_K_M · 2.8 of 46 GB
  13. 132x RTX 3090Graphics card · 48 GB~215 tok/s ModelFit est.Q4_K_M · 2.8 of 46 GB
  14. 14RTX 4080 SUPERGraphics card · 16 GB~213 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  15. 15RTX 4080Graphics card · 16 GB~209 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  16. 16RTX 5070Graphics card · 12 GB~205 tok/s ModelFit est.Q4_K_M · 2.8 of 11.0 GB
  17. 17RTX 4070 Ti SUPERGraphics card · 16 GB~199 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  18. 18M5 MaxMac · 48–128 GB · 3 variants~183 tok/s ModelFit est.Q4_K_M · 2.8 GB
  19. 19RX 7900 XTXGraphics card · 24 GB~179 tok/s ModelFit est.Q4_K_M · 2.8 of 23 GB
  20. 20M3 UltraMac · 96–512 GB · 5 variants~176 tok/s ModelFit est.Q4_K_M · 2.8 GB
  21. 212x RX 7900 XTXGraphics card · 48 GB~170 tok/s ModelFit est.Q4_K_M · 2.8 of 46 GB
  22. 22RX 7900 XTGraphics card · 20 GB~163 tok/s ModelFit est.Q4_K_M · 2.8 of 19.0 GB
  23. 23RTX 4070 TiGraphics card · 12 GB~160 tok/s ModelFit est.Q4_K_M · 2.8 of 11.0 GB
  24. 24RTX 4070 SUPERGraphics card · 12 GB~160 tok/s ModelFit est.Q4_K_M · 2.8 of 11.0 GB
  25. 25RTX 4070Graphics card · 12 GB~160 tok/s ModelFit est.Q4_K_M · 2.8 of 11.0 GB
  26. 26M4 MaxMac · 48–128 GB · 3 variants~154 tok/s ModelFit est.Q4_K_M · 2.8 GB
  27. 27RTX 5060Graphics card · 8 GB~151 tok/s ModelFit est.Q4_K_M · 2.8 of 7.0 GB
  28. 28RTX 5060 TiGraphics card · 16 GB~151 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  29. 29M2 UltraMac · 64–192 GB · 7 variants~151 tok/s ModelFit est.Q4_K_M · 2.8 GB
  30. 30M5 Max · 36 GB · 32cMac · 36 GB~145 tok/s ModelFit est.Q4_K_M · 2.8 of 24 GB
  31. 31RX 9070 XTGraphics card · 16 GB~144 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  32. 32RTX 3070Graphics card · 8 GB~134 tok/s ModelFit est.Q4_K_M · 2.8 of 7.0 GB
  33. 33RTX 3060 TiGraphics card · 8 GB~134 tok/s ModelFit est.Q4_K_M · 2.8 of 7.0 GB
  34. 34M1 UltraMac · 64–128 GB · 4 variants~134 tok/s ModelFit est.Q4_K_M · 2.8 GB
  35. 35M3 MaxMac · 48–128 GB · 3 variants~123 tok/s ModelFit est.Q4_K_M · 2.8 GB
  36. 36RTX 2070Graphics card · 8 GB~123 tok/s ModelFit est.Q4_K_M · 2.8 of 7.0 GB
  37. 37M4 Max · 36 GB · 32cMac · 36 GB~122 tok/s ModelFit est.Q4_K_M · 2.8 of 24 GB
  38. 38M5 ProMac · 24–64 GB · 6 variants~113 tok/s ModelFit est.Q4_K_M · 2.8 GB
  39. 39M2 MaxMac · 32–96 GB · 5 variants~113 tok/s ModelFit est.Q4_K_M · 2.8 GB
  40. 40RTX 3060Graphics card · 12 GB~112 tok/s ModelFit est.Q4_K_M · 2.8 of 11.0 GB
  41. 41M1 MaxMac · 32–64 GB · 4 variants~101 tok/s ModelFit est.Q4_K_M · 2.8 GB
  42. 42RTX 4060 TiGraphics card · 16 GB~99 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  43. 43M3 MaxMac · 36–96 GB · 2 variants~97 tok/s ModelFit est.Q4_K_M · 2.8 GB
  44. 44Tesla P100Graphics card · 16 GB~96 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  45. 45M4 ProMac · 24–64 GB · 6 variants~95 tok/s ModelFit est.Q4_K_M · 2.8 GB
  46. 46RTX 4060Graphics card · 8 GB~94 tok/s ModelFit est.Q4_K_M · 2.8 of 7.0 GB
  47. 47RX 9060 XTGraphics card · 16 GB~93 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  48. 48Tesla T4Graphics card · 16 GB~87 tok/s ModelFit est.Q4_K_M · 2.8 of 15.0 GB
  49. 49DGX Spark (GB10)AI mini PC · 128 GB~83 tok/s our est.Q4_K_M · 2.8 of 122 GB
  50. 50Ryzen AI Max+ 395 · 64 GBAI mini PC · 64 GB~78 tok/s our est.Q4_K_M · 2.8 of 48 GB
  51. 51Ryzen AI Max+ 395 · 128 GBAI mini PC · 128 GB~71 tok/s ModelFit est.Q4_K_M · 2.8 of 96 GB
  52. 52M2 ProMac · 16–32 GB · 2 variants~69 tok/s ModelFit est.Q4_K_M · 2.8 GB
  53. 53M6Mac · 24–32 GB · 3 variants~68 tok/s ModelFit est.Q4_K_M · 2.8 GB
  54. 54M1 ProMac · 16–32 GB · 2 variants~62 tok/s ModelFit est.Q4_K_M · 2.8 GB
  55. 55M6Mac · 16 GB · 2 variants~62 tok/s ModelFit est.Q4_K_M · 2.8 GB
  56. 56M5Mac · 16–32 GB · 6 variants~59 tok/s ModelFit est.Q4_K_M · 2.8 GB
  57. 57M3 ProMac · 18–36 GB · 2 variants~57 tok/s ModelFit est.Q4_K_M · 2.8 GB
  58. 58M4Mac · 16–32 GB · 3 variants~44 tok/s ModelFit est.Q4_K_M · 2.8 GB
  59. 59M3Mac · 8–24 GB · 3 variants~38 tok/s ModelFit est.Q4_K_M · 2.8 GB
  60. 60M2Mac · 8–24 GB · 3 variants~36 tok/s ModelFit est.Q4_K_M · 2.8 GB
  61. 61M1Mac · 8–16 GB · 2 variants~23 tok/s ModelFit est.Q4_K_M · 2.8 GB

Speeds are single-user decode estimates, not measurements. "ModelFit est." comes from ModelFit's model fitted on published runs (CC BY 4.0); "our est." is memory bandwidth divided by bytes read per token, times 0.55. Memory is what the best build needs, against what the machine can give a model. Method

About the model

Parameters
2B
Active per token
2B
Type
Dense
Good for
Chat

Builds and how to run them

QuantizationLoadMin. RAMCommand
Q4_K_M1.8 GB4 GBollama run gemma2:2b-instruct-q4_K_M

Load, minimum RAM and Ollama commands: ModelFit, CC BY 4.0. Load is memory with the model loaded, before long contexts.

Questions

Can I run Gemma 2 2B Instruct on a Mac?

Yes on a Mac with enough memory: the smallest Apple configuration listed here that runs it at 4-bit or better is the M3 · 8 GB.

How much memory does Gemma 2 2B Instruct need?

About 2.8 GB at the smallest build in our data, including a margin for the runtime and an 8K-token context. Longer contexts need more.

Does Gemma 2 2B Instruct run on a DGX Spark?

Yes, at Q4_K_M. The Spark has 128 GB of unified memory, about 121.6 GB usable.