Best local LLMs for the Rockchip RK3588 / RK3588S (8 GB)

Each list assumes the context that task needs, so a model that fits for chat may rank lower for long documents.

General chat at 8K context

  1. 57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.0 Tiny (8B-A1.3B) Q4_K_M · 24.4 tok/s · Runs well (tight fit)
  2. 58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking. Gemma 4 E2B Q4_K_M · 15.0 tok/s · Runs well
  3. 65Good quality (B) · 65 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. SenseNova Loop 9B Preview Q3_K_M · 4.6 tok/s · Slow (tight fit)
  4. 57Good quality (B) · 57 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Nanbeige4.2 3B Q3_K_M · 10.7 tok/s · Runs well
  5. 64Good quality (B) · 64 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Ornith 1.5 9B Q3_K_M · 4.6 tok/s · Slow (tight fit)

Coding at 32K context

  1. 57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking. MiniCPM5 2B Q4_K_M · 13.7 tok/s · Runs well
  2. 57Good quality (B) · 57 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Nemotron Labs Dory 1.0 4B Q3_K_M · 11.2 tok/s · Runs well
  3. 55Good quality (B) · 55 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking. Spark-X2.6 4B Q3_K_M · 10.9 tok/s · Runs well
  4. 58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Gemma 4 E2B Q4_K_M · 15.0 tok/s · Runs well
  5. 50Fair quality (C) · 50 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.0 Tiny (8B-A1.3B) Q3_K_M · 30.9 tok/s · Runs well (tight fit)

Reasoning at 32K context

  1. 57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking. MiniCPM5 2B Q4_K_M · 13.7 tok/s · Runs well
  2. 57Good quality (B) · 57 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Nemotron Labs Dory 1.0 4B Q3_K_M · 11.2 tok/s · Runs well
  3. 55Good quality (B) · 55 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking. Spark-X2.6 4B Q3_K_M · 10.9 tok/s · Runs well
  4. 58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Gemma 4 E2B Q4_K_M · 15.0 tok/s · Runs well
  5. 50Fair quality (C) · 50 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.0 Tiny (8B-A1.3B) Q3_K_M · 30.9 tok/s · Runs well (tight fit)

Documents (RAG) at 32K context

  1. 58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking. Gemma 4 E2B Q4_K_M · 15.0 tok/s · Runs well
  2. 41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Unlimited OCR 2 (3B-A0.6B) Q4_K_M · 48.3 tok/s · Runs great
  3. 38Basic quality (D) · 38 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking. Unlimited OCR (3B-A0.6B) Q4_K_M · 48.3 tok/s · Runs great
  4. 36Basic quality (D) · 36 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Granite 4.0 H Tiny (7B-A1B) Q3_K_M · 39.0 tok/s · Runs great
  5. 51Fair quality (C) · 51 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. TMax 4B Q3_K_M · 10.6 tok/s · Runs well

Agents & tools at 64K context

  1. 57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. MiniCPM5 2B Q4_K_M · 13.7 tok/s · Runs well (tight fit)
  2. 55Good quality (B) · 55 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Spark-X2.6 4B Q3_K_M · 10.9 tok/s · Runs well (tight fit)
  3. 36Basic quality (D) · 36 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Granite 4.0 H Tiny (7B-A1B) Q3_K_M · 39.0 tok/s · Runs great
  4. 43Fair quality (C) · 43 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. LFM2.5 8B-A1B Q3_K_M · 27.2 tok/s · Runs well (tight fit)
  5. 51Fair quality (C) · 51 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. TMax 4B Q3_K_M · 10.6 tok/s · Runs well

Reverse engineering at 32K context

  1. 58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Gemma 4 E2B Q4_K_M · 15.0 tok/s · Runs well
  2. 50Fair quality (C) · 50 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.0 Tiny (8B-A1.3B) Q3_K_M · 30.9 tok/s · Runs well (tight fit)
  3. 57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. MiniCPM5 2B Q4_K_M · 13.7 tok/s · Runs well
  4. 41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. Unlimited OCR 2 (3B-A0.6B) Q4_K_M · 48.3 tok/s · Runs great
  5. 57Good quality (B) · 57 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. Nemotron Labs Dory 1.0 4B Q3_K_M · 11.2 tok/s · Runs well

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.

Was this page helpful?