AMD Strix Point (Radeon 890M)
Other42
Run records
14
Models covered
4
Frameworks covered
L2 Cross-framework verified × 42
Evidence mix
Typical measured performance
Published benchmark averages per model × framework × quantization on this hardware.
- llama.cpp · Q5_K_M Decode 58.41 · Prefill 300.39 ·
- Ollama · Q3_K_L Decode 78.29 · Prefill 430.35 ·
- Ollama · Q8_0 Decode 43.81 · Prefill 179.92 ·
- vLLM · IQ3_S Decode 65.94 · Prefill 339.59 ·
- Ollama · IQ3_XS Decode 62.85 · Prefill 248.56 ·
- Ollama · IQ4_XS Decode 46.77 · Prefill 176.41 ·
- vLLM · IQ3_XS Decode 172.31 · Prefill 798.79 ·
- Strata · Q5_K_M Decode 163.87 · Prefill 897.86 ·
- vLLM · Q2_K Decode 74.53 · Prefill 292.59 ·
- vLLM · Q3_K_M Decode 62.84 · Prefill 312.12 ·
- Strata · Q4_K_M Decode 68.09 · Prefill 270.12 ·
- vLLM · Q4_K_M Decode 56.8 · Prefill 217.7 ·
- Strata · IQ1_S Decode 121.73 · Prefill 484.23 ·
- Strata · Q5_K_S Decode 66.77 · Prefill 312.73 ·
- vLLM · FP16 Decode 36.08 · Prefill 186.22 ·
- vLLM · IQ2_XS Decode 112.83 · Prefill 404.8 ·
- Ollama · IQ4_XS Decode 68.56 · Prefill 288.08 ·
- Ollama · Q3_K_L Decode 75.1 · Prefill 314.07 ·
- llama.cpp · Q8_0 Decode 68.66 · Prefill 239.32 ·
- Strata · Q3_K_L Decode 154.95 · Prefill 755.86 ·
- Strata · Q4_K_M Decode 149.78 · Prefill 509.73 ·
- llama.cpp · IQ4_XS Decode 75.65 · Prefill 261.83 ·
- vLLM · Q5_K_S Decode 60.14 · Prefill 291.94 ·
- vLLM · IQ1_S Decode 79.93 · Prefill 394.12 ·
- vLLM · Q8_0 Decode 38.46 · Prefill 182.92 ·
- Ollama · Q5_K_M Decode 41.64 · Prefill 184.88 ·
- llama.cpp · Q4_K_S Decode 160.41 · Prefill 865.77 ·
- Strata · Q4_K_S Decode 231.08 · Prefill 1081.57 ·
- Ollama · IQ2_M Decode 98.85 · Prefill 414.8 ·
- Ollama · Q6_K Decode 60.33 · Prefill 317.9 ·
- Strata · IQ3_S Decode 128.16 · Prefill 598.86 ·
- vLLM · Q4_K_S Decode 178.08 · Prefill 918.71 ·
- vLLM · Q8_0 Decode 112.43 · Prefill 505.72 ·
- Strata · IQ2_M Decode 281.13 · Prefill 1305.8 ·
- Strata · IQ3_XS Decode 273.43 · Prefill 1002.15 ·
- llama.cpp · IQ1_S Decode 145.7 · Prefill 692.08 ·
- llama.cpp · IQ4_XS Decode 106.04 · Prefill 563.55 ·
- Ollama · FP16 Decode 47.28 · Prefill 250.08 ·
- Ollama · Q8_0 Decode 57.25 · Prefill 264.02 ·
- vLLM · IQ3_XS Decode 93.71 · Prefill 479.79 ·
- vLLM · Q6_K Decode 56.27 · Prefill 255.06 ·
- Ollama · Q2_K Decode 85.31 · Prefill 404.47 ·
Fit check
Physics-based estimate from VRAM and memory bandwidth: pick a model to see which quantizations fit and the expected speed
VRAM is not recorded for this hardware; fit cannot be computed
Formula: tok/s = (bandwidth ÷ model size) × 0.55, adapted from llmfit; see run records for measured numbers
Key specs
- Vendor
- AMD
- Memory bandwidth
- 153 GB/s
Run records
42 configurations
| Model | Framework | Quantization | Decode | Gen time | Samples |
|---|---|---|---|---|---|
| Qwen3.5-4B | llama.cpp | Q5_K_M | 58.4 tok/s | — | 1 |
| Qwen3.5-4B | Ollama | Q3_K_L | 78.3 tok/s | — | 1 |
| Qwen3.5-4B | Ollama | Q8_0 | 43.8 tok/s | — | 1 |
| Qwen3.5-9B | vLLM | IQ3_S | 65.9 tok/s | — | 1 |
| Qwen3.5-9B | Ollama | IQ3_XS | 62.9 tok/s | — | 1 |
| Qwen3.5-9B | Ollama | IQ4_XS | 46.8 tok/s | — | 1 |
| Qwen3.5-0.8B | vLLM | IQ3_XS | 172.3 tok/s | — | 1 |
| Qwen3.5-0.8B | Strata | Q5_K_M | 163.9 tok/s | — | 1 |
| Qwen3.8-9B | vLLM | Q2_K | 74.5 tok/s | — | 1 |
| Qwen3.8-9B | vLLM | Q3_K_M | 62.8 tok/s | — | 1 |
| Qwen3.8-9B | Strata | Q4_K_M | 68.1 tok/s | — | 1 |
| Qwen3-8B | vLLM | Q4_K_M | 56.8 tok/s | — | 1 |
| Qwen3-8B | Strata | IQ1_S | 121.7 tok/s | — | 1 |
| Qwen3-8B | Strata | Q5_K_S | 66.8 tok/s | — | 1 |
| Qwen3-4B | vLLM | FP16 | 36.1 tok/s | — | 1 |
| Qwen3-4B | vLLM | IQ2_XS | 112.8 tok/s | — | 1 |
| Qwen3-4B | Ollama | IQ4_XS | 68.6 tok/s | — | 1 |
| Qwen3-4B | Ollama | Q3_K_L | 75.1 tok/s | — | 1 |
| Qwen3-1.7B | llama.cpp | Q8_0 | 68.7 tok/s | — | 1 |
| Qwen3-1.7B | Strata | Q3_K_L | 154.9 tok/s | — | 1 |
| Qwen3-1.7B | Strata | Q4_K_M | 149.8 tok/s | — | 1 |
| Qwen3-4B-Thinking-2507 | llama.cpp | IQ4_XS | 75.7 tok/s | — | 1 |
| Qwen3-4B-Thinking-2507 | vLLM | Q5_K_S | 60.1 tok/s | — | 1 |
| Qwen3-Coder-Next | vLLM | IQ1_S | 79.9 tok/s | — | 1 |
| Qwen3-Coder-Next | vLLM | Q8_0 | 38.5 tok/s | — | 1 |
| Qwen3-Coder-Next | Ollama | Q5_K_M | 41.6 tok/s | — | 1 |
| Qwen2.5-0.5B-Instruct | llama.cpp | Q4_K_S | 160.4 tok/s | — | 1 |
| Qwen2.5-0.5B-Instruct | Strata | Q4_K_S | 231.1 tok/s | — | 1 |
| Qwen2.5-Omni-3B | Ollama | IQ2_M | 98.8 tok/s | — | 1 |
| Qwen2.5-Omni-3B | Ollama | Q6_K | 60.3 tok/s | — | 1 |
| Qwen2.5-Omni-3B | Strata | IQ3_S | 128.2 tok/s | — | 1 |
| Qwen3-Embedding-0.6B | vLLM | Q4_K_S | 178.1 tok/s | — | 1 |
| Qwen3-Embedding-0.6B | vLLM | Q8_0 | 112.4 tok/s | — | 1 |
| Qwen3-Embedding-0.6B | Strata | IQ2_M | 281.1 tok/s | — | 1 |
| Qwen3-Embedding-0.6B | Strata | IQ3_XS | 273.4 tok/s | — | 1 |
| gemma-4-E2B-it | llama.cpp | IQ1_S | 145.7 tok/s | — | 1 |
| gemma-4-E2B-it | llama.cpp | IQ4_XS | 106 tok/s | — | 1 |
| gemma-4-E2B-it | Ollama | FP16 | 47.3 tok/s | — | 1 |
| gemma-4-E2B-it | Ollama | Q8_0 | 57.3 tok/s | — | 1 |
| gemma-4-E4B-it | vLLM | IQ3_XS | 93.7 tok/s | — | 1 |
| gemma-4-E4B-it | vLLM | Q6_K | 56.3 tok/s | — | 1 |
| gemma-4-E4B-it | Ollama | Q2_K | 85.3 tok/s | — | 1 |