Best local LLMs for the AMD EPYC 7002 / 7003 series NAS (512 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. 96Excellent quality (A+) · 96 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3.1 Flash (428B-A23B) Q6_K · 4.2 tok/s · Slow
  2. 92Excellent quality (A+) · 92 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.1 Flash 1019 (560B-A25B) Q4_K_M · 5.5 tok/s · Slow
  3. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. K2-Horizon (375B-A23B) Q6_K · 4.2 tok/s · Slow
  4. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3 (427B) Q6_K · 3.8 tok/s · Slow
  5. 91Excellent quality (A+) · 91 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. Beam 1.0 (501B-A23B) Q5_K_M · 4.8 tok/s · Slow

Coding at 32K context

  1. 96Excellent quality (A+) · 96 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3.1 Flash (428B-A23B) Q6_K · 4.2 tok/s · Slow
  2. 92Excellent quality (A+) · 92 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.1 Flash 1019 (560B-A25B) Q4_K_M · 5.5 tok/s · Slow
  3. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. K2-Horizon (375B-A23B) Q6_K · 4.2 tok/s · Slow
  4. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3 (427B) Q6_K · 3.8 tok/s · Slow
  5. 91Excellent quality (A+) · 91 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. Beam 1.0 (501B-A23B) Q5_K_M · 4.8 tok/s · Slow

Reasoning at 32K context

  1. 96Excellent quality (A+) · 96 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3.1 Flash (428B-A23B) Q6_K · 4.2 tok/s · Slow
  2. 92Excellent quality (A+) · 92 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.1 Flash 1019 (560B-A25B) Q4_K_M · 5.5 tok/s · Slow
  3. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. K2-Horizon (375B-A23B) Q6_K · 4.2 tok/s · Slow
  4. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3 (427B) Q6_K · 3.8 tok/s · Slow
  5. 91Excellent quality (A+) · 91 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. Beam 1.0 (501B-A23B) Q5_K_M · 4.8 tok/s · Slow

Documents (RAG) at 32K context

  1. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3 (427B) Q6_K · 3.8 tok/s · Slow
  2. 90Excellent quality (A+) · 90 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking. GLM-4.7 (355B-A32B) Q5_K_M · 3.6 tok/s · Slow
  3. 91Excellent quality (A+) · 91 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking. Qwen3.8 Flash Next (125B-A6B) Q4_K_M · 21.6 tok/s · Runs well
  4. 88Excellent quality (A+) · 88 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking. Qwen3 Coder 480B-A35B Q5_K_M · 3.3 tok/s · Slow
  5. 90Excellent quality (A+) · 90 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M2.7 (230B-A10B) Q5_K_M · 11.1 tok/s · Runs well

Agents & tools at 64K context

  1. 96Excellent quality (A+) · 96 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3.1 Flash (428B-A23B) Q6_K · 4.2 tok/s · Slow
  2. 92Excellent quality (A+) · 92 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.1 Flash 1019 (560B-A25B) Q4_K_M · 5.5 tok/s · Slow
  3. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. K2-Horizon (375B-A23B) Q6_K · 4.2 tok/s · Slow
  4. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3 (427B) Q6_K · 3.8 tok/s · Slow
  5. 91Excellent quality (A+) · 91 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. Beam 1.0 (501B-A23B) Q5_K_M · 4.8 tok/s · Slow

Reverse engineering at 32K context

  1. 96Excellent quality (A+) · 96 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking. MiniMax M3.1 Flash (428B-A23B) Q6_K · 4.2 tok/s · Slow
  2. 87Excellent quality (A+) · 87 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. DeepSeek R1 (671B) Q4_K_M · 3.7 tok/s · Slow (tight fit)
  3. 88Excellent quality (A+) · 88 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. Qwen3.5 35B-A3B Q4_K_M · 40.2 tok/s · Runs great
  4. 92Excellent quality (A+) · 92 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. Ling 3.1 Flash 1019 (560B-A25B) Q4_K_M · 5.5 tok/s · Slow
  5. 93Excellent quality (A+) · 93 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. K2-Horizon (375B-A23B) Q6_K · 4.2 tok/s · Slow

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?