| Model | Verdict | Decode tok/s | Max ctx | Size | Chat | Coding | |
|---|---|---|---|---|---|---|---|
| Qwen/Qwen3-30B-A3B-Instruct-2507 lmstudio-community/Qwen3-30B-A3B-Instruct-2507-MLX-8bit · MLX-8BIT | Fits | 68.827.5–75.7 | 256K | 30.5B | p85 | – | |
| Qwen/Qwen3.5-122B-A10B unsloth/Qwen3.5-122B-A10B-GGUF · Q5_K_S | Fits | 64.725.9–71.2 | 256K | 125.1B | p77 | p49 | |
| Qwen/Qwen3.8-Flash-Next ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF · IQ3_S | Fits | 81.832.7–90.0 | 256K | 180B | p68 | p51 | |
| google/gemma-4-31B-it lmstudio-community/gemma-4-31B-it-MLX-8bit · MLX-8BIT | Fits | 14.05.60–15.4 | 256K | 31.3B | p63 | – | |
| MiniMaxAI/MiniMax-M2.5 unsloth/MiniMax-M2.5-GGUF · Q2_K_L | Fits | 49.219.7–54.1 | 131K | 228.7B | p46 | p61 | |
| Qwen/Qwen3.5-27B unsloth/Qwen3.5-27B-GGUF · Q8_0 | Fits | 17.46.90–19.1 | 256K | 27.8B | p60 | p46 | |
| Qwen/Qwen3.8-27B lmstudio-community/Qwen3.8-27B-MLX-8bit · MLX-8BIT | Fits | 17.16.80–18.8 | 256K | 27.8B | p59 | p33 | |
| Qwen/Qwen3.6-27B lmstudio-community/Qwen3.6-27B-MLX-8bit · MLX-8BIT | Fits | 17.16.80–18.8 | 256K | 27.8B | p51 | p54 | |
| nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 unsloth/NVIDIA-Nemotron-3-Super-120B-A12B-GGUF · MXFP4_MOE | Fits | 76.430.6–84.0 | 256K | 123.6B | p50 | p15 | |
| Qwen/Qwen3.5-35B-A3B unsloth/Qwen3.5-35B-A3B-GGUF · Q8_0 | Fits | 96.438.6–106 | 256K | 36B | p49 | p35 | |
| google/gemma-4-26B-A4B-it lmstudio-community/gemma-4-26B-A4B-it-MLX-8bit · MLX-8BIT | Fits | 79.331.7–87.2 | 256K | 25.8B | p49 | – | |
| openai/gpt-oss-120b unsloth/gpt-oss-120b-GGUF · F16 | Fits | 87.935.2–96.7 | 128K | 117B | p47 | p15 | |
| Qwen/Qwen3.5-9B lmstudio-community/Qwen3.5-9B-MLX-8bit · MLX-8BIT | Fits | 43.917.6–48.3 | 256K | 9.7B | p46 | – | |
| ornith-ai/Ornith-1.5-9B BF16 | Fits | 26.910.7–29.5 | 256K | 9.7B | p46 | p24 | |
| Qwen/Qwen3.6-35B-A3B lmstudio-community/Qwen3.6-35B-A3B-MLX-8bit · MLX-8BIT | Fits | 95.538.2–105 | 256K | 36B | p43 | p41 | |
| openbmb/MiniCPM-V-4.7-35B-A3B openbmb/MiniCPM-V-4.7-35B-A3B-gguf · Q8_0 | Fits | 96.638.6–106 | 256K | 35.2B | p42 | – | |
| ornith-ai/Ornith-1.0-35B lmstudio-community/Ornith-1.0-35B-MLX-8bit · MLX-8BIT | Fits | 95.738.3–105 | 256K | 35.1B | – | p40 | |
| openai/gpt-oss-20b MXFP4 | Fits | 69.327.7–76.3 | 128K | 21.1B | p39 | p15 | |
| Qwen/Qwen3.5-4B BF16 | Fits | 44.017.6–48.4 | 256K | 4.7B | p39 | – | |
| google/gemma-4-12B-it unsloth/gemma-4-12b-it-GGUF · Q8_0 | Fits | 32.727.8–37.6 | 256K | 12B | p38 | – | |
| Qwen/Qwen3-4B-Instruct-2507 BF16 | Fits | 41.535.2–47.7 | 256K | 4B | p38 | – | |
| nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 BF16 | Fits | 93.837.5–103 | 256K | 32.9B | p37 | p9 | |
| nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 MIXED_PRECISION | Fits | 11144.5–122 | 1M | 47.1B | p36 | p9 | |
| poolside/Laguna-S-2.1 TensorFold/Laguna-S-2.1-MLX-4bit · MLX-4BIT | Fits | 72.328.9–79.6 | 697K | 117.6B | – | p34 | |
| nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 BF16 | Fits | 95.938.4–106 | 256K | 31.6B | p33 | – | |
| poolside/Laguna-XS-2.1 AtomicChat/Laguna-XS-2.1-MLX-8bit · MLX-8BIT | Fits | 98.239.3–108 | 256K | 33.4B | – | p30 | |
| Qwen/Qwen2.5-7B BF16 | Fits | 30.626.0–35.1 | 128K | 7.6B | p29 | – | |
| google/gemma-4-E4B-it lmstudio-community/gemma-4-E4B-it-MLX-8bit · MLX-8BIT | Fits | 60.824.3–66.9 | 128K | 8B | p27 | – | |
| openbmb/MiniCPM5-2B BF16 | Fits | 65.855.9–75.7 | 128K | 2.5B | p27 | p3 | |
| poolside/Laguna-XS.2 BF16 | Fits | 61.124.4–67.2 | 256K | 33.4B | – | p25 | |
| Qwen/Qwen3.5-2B BF16 | Fits | 74.829.9–82.3 | 256K | 2.3B | p24 | – | |
| ornith-ai/Ornith-1.0-9B lmstudio-community/Ornith-1.0-9B-MLX-8bit · MLX-8BIT | Fits | 44.017.6–48.5 | 256K | 9.4B | – | p20 | |
| google/gemma-4-E2B-it lmstudio-community/gemma-4-E2B-it-MLX-8bit · MLX-8BIT | Fits | 85.934.4–94.5 | 128K | 5.1B | p18 | – | |
| Qwen/Qwen2.5-1.5B BF16 | Fits | 86.173.2–99.0 | 128K | 1.5B | p15 | – | |
| mistralai/Mistral-7B-Instruct-v0.2 BF16 | Fits | 28.724.4–33.0 | 32K | 7.2B | p14 | – | |
| Qwen/Qwen2.5-0.5B BF16 | Fits | 143121–164 | 32K | 0.5B | p5 | – | |
| Qwen/Qwen3.5-0.8B BF16 | Fits | 12148.4–133 | 256K | 0.9B | p4 | – | |
| Qwen/Qwen2.5-0.5B-Instruct BF16 | Fits | 143121–164 | 32K | 0.5B | – | – | |
| incoai/Qwen3.8-27B-DFlash2 BF16 | Fits | 143121–164 | 256K | 1.9B | – | – | |
| datalab-to/surya-ocr-2 BF16 | Fits | 13252.8–145 | 256K | 0.7B | – | – | |
| Qwen/Qwen3.5-0.8B-Base BF16 | Fits | 12148.4–133 | 256K | 0.9B | – | – | |
| baidu/Unlimited-OCR-2 AtomicChat/Unlimited-OCR-2-GGUF · BF16 | Fits | 11646.6–128 | 128K | 3.3B | – | – | |
| nvidia/Qwen3.6-35B-A3B-NVFP4 MIXED_PRECISION | Fits | 11646.3–127 | 256K | 34B | – | – | |
| ornith-ai/Ornith-1.5-35B-A3B-NVFP4 MIXED_PRECISION | Fits | 11646.3–127 | 256K | 34B | – | – | |
| openbmb/MiniCPM-V-4.7-1B BF16 | Fits | 10240.7–112 | 256K | 1.6B | – | – | |
| Qwen/Qwen3-0.6B-Base BF16 | Fits | 96.882.2–111 | 32K | 0.6B | – | – | |
| Qwen/Qwen3-0.6B BF16 | Fits | 91.778.0–106 | 40K | 0.8B | – | – | |
| ornith-ai/Ornith-1.0-35B-GGUF Q8_0 | Fits | 86.534.6–95.2 | 256K | 34.7B | – | – | |
| Qwen/Qwen2.5-1.5B-Instruct BF16 | Fits | 86.173.2–99.0 | 32K | 1.5B | – | – | |
| ornith-ai/Ornith-1.5-35B-A3B-GGUF Q8_0 | Fits | 85.634.2–94.2 | 256K | 35.5B | – | – | |
| nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF · Q8_0 | Fits | 83.233.3–91.5 | 256K | 4B | – | – | |
| Qwen/Qwen2.5-3B-Instruct Qwen/Qwen2.5-3B-Instruct-GGUF · Q8_0 | Fits | 75.564.2–86.9 | 32K | 3.1B | – | – | |
| Qwen/Qwen2-VL-2B-Instruct BF16 | Fits | 71.260.5–81.9 | 32K | 2.2B | – | – | |
| ornith-ai/Ornith-1.5-35B-A3B-FP8 COMPRESSED-TENSORS | Fits | 70.628.2–77.6 | 256K | 36B | – | – | |
| Qwen/Qwen3-1.7B-Base BF16 | Fits | 69.358.9–79.7 | 32K | 1.7B | – | – | |
| Qwen/Qwen3-VL-2B-Instruct Qwen/Qwen3-VL-2B-Instruct-GGUF · F16 | Fits | 69.258.8–79.6 | 256K | 2.1B | – | – | |
| Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8 FP8 | Fits | 69.127.6–76.0 | 256K | 30.5B | – | – | |
| baidu/Unlimited-OCR BF16 | Fits | 67.827.1–74.6 | 32K | 3.3B | – | – | |
| openbmb/MiniCPM5-2B-GGUF F16 | Fits | 65.855.9–75.6 | 128K | 2.5B | – | – | |
| Qwen/Qwen3-1.7B BF16 | Fits | 64.254.6–73.8 | 40K | 2B | – | – | |
| prism-ml/Bonsai-27B-mlx-1bit MLX-1BIT | Fits | 62.224.9–68.5 | 256K | 26.4B | – | – | |
| Qwen/Qwen2.5-VL-3B-Instruct unsloth/Qwen2.5-VL-3B-Instruct-GGUF · BF16 | Fits | 55.847.5–64.2 | 125K | 3.8B | – | – | |
| Qwen/Qwen3-VL-4B-Instruct lmstudio-community/Qwen3-VL-4B-Instruct-MLX-8bit · MLX-8BIT | Fits | 53.321.3–58.6 | 256K | 4.4B | – | – | |
| ornith-ai/Ornith-1.5-122B-GGUF Q5_K_M | Fits | 52.521.0–57.7 | 256K | 124.6B | – | – | |
| Qwen/Qwen3-Coder-30B-A3B-Instruct BF16 | Fits | 50.220.1–55.3 | 256K | 30.5B | – | – | |
| Qwen/Qwen3-30B-A3B BF16 | Fits | 50.220.1–55.3 | 40K | 30.5B | – | – | |
| ornith-ai/Ornith-1.5-9B-NVFP4 MIXED_PRECISION | Fits | 50.020.0–55.0 | 256K | 9B | – | – | |
| Qwen/Qwen2.5-7B-Instruct Qwen/Qwen2.5-7B-Instruct-GGUF · Q8_0 | Fits | 48.441.2–55.7 | 32K | 7.6B | – | – | |
| Qwen/Qwen2.5-VL-7B-Instruct unsloth/Qwen2.5-VL-7B-Instruct-GGUF · Q8_0 | Fits | 48.241.0–55.5 | 125K | 8.3B | – | – | |
| Qwen/Qwen2-VL-7B-Instruct ggml-org/Qwen2-VL-7B-Instruct-GGUF · Q8_0 | Fits | 48.241.0–55.5 | 32K | 8.3B | – | – | |
| ornith-ai/Ornith-1.0-9B-GGUF Q8_0 | Fits | 46.818.7–51.4 | 256K | 9B | – | – | |
| prism-ml/Ternary-Bonsai-27B-mlx-2bit MLX-2BIT | Fits | 45.718.3–50.2 | 256K | 26.4B | – | – | |
| ornith-ai/Ornith-1.5-9B-GGUF Q8_0 | Fits | 45.718.3–50.3 | 256K | 9.2B | – | – | |
| Qwen/Qwen3.5-4B-Base BF16 | Fits | 44.017.6–48.4 | 256K | 4.7B | – | – | |
| Qwen/Qwen2.5-Coder-7B-Instruct Qwen/Qwen2.5-Coder-7B-Instruct-GGUF · Q4_K_M | Fits | 43.937.3–50.5 | 32K | 7.6B | – | – | |
| HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive Q8_K_P | Fits | 43.617.4–48.0 | 128K | 7.5B | – | – | |
| Qwen/Qwen3-8B-Base Qwen/Qwen3-Embedding-8B-GGUF · Q8_0 | Fits | 43.536.9–50.0 | 32K | 8.2B | – | – | |
| Qwen/Qwen3-8B lmstudio-community/Qwen3-8B-MLX-8bit · MLX-8BIT | Fits | 41.516.6–45.6 | 40K | 8.2B | – | – | |
| Qwen/Qwen3-4B BF16 | Fits | 41.535.2–47.7 | 40K | 4B | – | – | |
| datalab-to/chandra-ocr-2 BF16 | Fits | 39.916.0–43.9 | 256K | 5.3B | – | – | |
| Qwen/Qwen3-VL-8B-Instruct lmstudio-community/Qwen3-VL-8B-Instruct-MLX-8bit · MLX-8BIT | Fits | 38.215.3–42.1 | 256K | 8.8B | – | – | |
| Qwen/Qwen3-14B lmstudio-community/Qwen3-14B-MLX-8bit · MLX-8BIT | Fits | 26.910.7–29.5 | 40K | 14.8B | – | – | |
| Qwen/Qwen2.5-14B-Instruct Qwen/Qwen2.5-14B-Instruct-GGUF · Q8_0 | Fits | 26.122.1–30.0 | 32K | 14.8B | – | – | |
| Qwen/Qwen2.5-Coder-14B-Instruct Qwen/Qwen2.5-Coder-14B-Instruct-GGUF · Q4_K_M | Fits | 23.620.1–27.2 | 32K | 14.8B | – | – | |
| RadixArk/Qwen3.8-27B-NVFP4 MIXED_PRECISION | Fits | 22.18.80–24.3 | 256K | 26.4B | – | – | |
| Qwen/Qwen2.5-32B-Instruct lmstudio-community/Qwen2.5-32B-Instruct-MLX-8bit · MLX-8BIT | Fits | 13.35.30–14.6 | 32K | 32.8B | – | – | |
| Qwen/Qwen2.5-Coder-32B-Instruct lmstudio-community/Qwen2.5-Coder-32B-Instruct-MLX-8bit · MLX-8BIT | Fits | 13.35.30–14.6 | 32K | 32.8B | – | – | |
| Qwen/Qwen3-32B lmstudio-community/Qwen3-32B-MLX-8bit · MLX-8BIT | Fits | 13.15.30–14.4 | 40K | 32.8B | – | – | |
| farbodtavakkoli/OTel-2.0-LLM-31B-IT wfakhri/OTel-2.0-LLM-31B-IT-GGUF · Q8_0 | Fits | 13.111.2–15.1 | 256K | 31.3B | – | – | |
| nvidia/Gemma-4-31B-IT-NVFP4 NVFP4 | Fits | – | 256K | 29.2B | – | – | |
| nvidia/Gemma-4-26B-A4B-NVFP4 NVFP4 | Fits | – | 256K | 24.5B | – | – | |
| nvidia/Qwen2.5-VL-7B-Instruct-NVFP4 NVFP4 | Fits | – | 125K | 8.2B | – | – |
Quality = benchmark percentile against all indexed models (p50 = median). Speeds are single-user decode ranges.