| Model | Verdict | Decode tok/s | Max ctx | Size | Chat | Coding | |
|---|---|---|---|---|---|---|---|
| Qwen/Qwen3-30B-A3B-Instruct-2507 BF16 | Fits | 17670.5–194 | 256K | 30.5B | p85 | – | |
| google/gemma-4-31B-it unsloth/gemma-4-31B-it-GGUF · Q8_0 | Fits | 78.431.4–98.1 | 256K | 31.3B | p62 | – | |
| Qwen/Qwen3.5-27B unsloth/Qwen3.5-27B-GGUF · Q8_0 | Fits | 99.639.8–110 | 256K | 27.8B | p60 | p46 | |
| Qwen/Qwen3.8-27B unsloth/Qwen3.8-27B-GGUF · Q8_0 | Fits | 98.739.5–109 | 256K | 27.8B | p58 | p31 | |
| Qwen/Qwen3.6-27B unsloth/Qwen3.6-27B-GGUF · Q8_0 | Fits | 99.639.8–110 | 256K | 27.8B | p50 | p54 | |
| Qwen/Qwen3.5-35B-A3B BF16 | Fits | 17168.2–188 | 256K | 36B | p48 | p35 | |
| google/gemma-4-26B-A4B-it BF16 | Fits | 16766.9–184 | 256K | 25.8B | p48 | – | |
| Qwen/Qwen3.5-9B BF16 | Fits | 13051.8–143 | 256K | 9.7B | p46 | – | |
| ornith-ai/Ornith-1.5-9B BF16 | Fits | 13051.8–143 | 256K | 9.7B | p44 | p23 | |
| Qwen/Qwen3.6-35B-A3B BF16 | Fits | 17168.2–188 | 256K | 36B | p43 | p41 | |
| openbmb/MiniCPM-V-4.7-35B-A3B BF16 | Fits | 18072.0–198 | 256K | 35.2B | p42 | – | |
| openai/gpt-oss-20b MXFP4 | Fits | 19578.1–215 | 128K | 21.1B | p39 | p15 | |
| Qwen/Qwen3.5-4B BF16 | Fits | 16566.0–181 | 256K | 4.7B | p39 | – | |
| ornith-ai/Ornith-1.0-35B BF16 | Fits | 18172.6–200 | 256K | 35.1B | – | p38 | |
| google/gemma-4-12B-it BF16 | Fits | 10943.5–120 | 256K | 12B | p38 | – | |
| Qwen/Qwen3-4B-Instruct-2507 BF16 | Fits | 16164.5–177 | 256K | 4B | p38 | – | |
| nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 unsloth/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-GGUF · BF16 | Fits | 10943.8–137 | 256K | 32.9B | p37 | p8 | |
| nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 MIXED_PRECISION | Fits | 22389.1–245 | 1M | 47.1B | p36 | p8 | |
| nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 unsloth/Nemotron-3-Nano-30B-A3B-GGUF · BF16 | Fits | 11144.3–138 | 256K | 31.6B | p33 | – | |
| poolside/Laguna-XS-2.1 BF16 | Fits | 18775.0–206 | 256K | 33.4B | – | p30 | |
| poolside/Laguna-XS.2 BF16 | Fits | 18775.0–206 | 256K | 33.4B | – | p25 | |
| ornith-ai/Ornith-1.0-9B BF16 | Fits | 13152.5–144 | 256K | 9.4B | – | p19 | |
| ornith-ai/Ornith-1.5-35B-A3B-NVFP4 MIXED_PRECISION | Fits | 22690.2–248 | 256K | 34B | – | – | |
| nvidia/Qwen3.6-35B-A3B-NVFP4 MIXED_PRECISION | Fits | 22690.2–248 | 256K | 34B | – | – | |
| Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8 FP8 | Fits | 19879.2–218 | 256K | 30.5B | – | – | |
| prism-ml/Bonsai-27B-mlx-1bit MLX-1BIT | Fits | 19075.8–209 | 256K | 26.4B | – | – | |
| Qwen/Qwen3-30B-A3B BF16 | Fits | 17670.5–194 | 40K | 30.5B | – | – | |
| Qwen/Qwen3-Coder-30B-A3B-Instruct BF16 | Fits | 17670.5–194 | 256K | 30.5B | – | – | |
| prism-ml/Ternary-Bonsai-27B-mlx-2bit MLX-2BIT | Fits | 16867.3–185 | 256K | 26.4B | – | – | |
| ornith-ai/Ornith-1.5-9B-GGUF BF16 | Fits | 13252.6–145 | 256K | 9.2B | – | – | |
| RadixArk/Qwen3.8-27B-NVFP4 MIXED_PRECISION | Fits | 11646.5–128 | 256K | 26.4B | – | – | |
| ornith-ai/Ornith-1.0-35B-GGUF BF16 | Fits | 11044.0–138 | 256K | 34.7B | – | – | |
| ornith-ai/Ornith-1.5-35B-A3B-GGUF BF16 | Fits | 10943.8–137 | 256K | 35.5B | – | – | |
| Qwen/Qwen3-14B BF16 | Fits | 95.938.4–106 | 40K | 14.8B | – | – | |
| Qwen/Qwen2.5-14B-Instruct BF16 | Fits | 94.837.9–104 | 32K | 14.8B | – | – | |
| Qwen/Qwen2.5-Coder-14B-Instruct BF16 | Fits | 94.837.9–104 | 32K | 14.8B | – | – | |
| Qwen/Qwen2.5-32B-Instruct Qwen/Qwen2.5-32B-Instruct-GGUF · Q8_0 | Fits | 73.729.5–92.2 | 32K | 32.8B | – | – | |
| Qwen/Qwen2.5-Coder-32B-Instruct Qwen/Qwen2.5-Coder-32B-Instruct-GGUF · Q4_K_M | Fits | 68.127.3–85.2 | 32K | 32.8B | – | – | |
| Qwen/Qwen3-32B BF16 | Fits | 54.621.8–60.1 | 40K | 32.8B | – | – | |
| farbodtavakkoli/OTel-2.0-LLM-31B-IT BF16 | Fits | 32.312.9–35.6 | 256K | 31.3B | – | – |
Quality = benchmark percentile against all indexed models (p50 = median). Speeds are single-user decode ranges.