Files and memory at 8K context
No verified quantized file for the requested formats.
No verified file: Q4_K_M, Q5_K_M, Q8_0, FP16, BF16, MLX 4-bit, MLX 8-bit. Sizes are exact file bytes (text weights; projectors listed separately). “+” = recurrent state or other terms unknown. Smallest fit adds 0.5 GiB spare; Macs use 75% of memory.
Quality
Artificial Analysis Intelligence Index
43.6
Variant: Kimi K3 (Max) · observed 2026-10-17 · source
Rank #4 of 75 researched 2026 models.
Arena text rating
1,475
95% CI 1470–1479 · 32,795 votes · published 2026-10-15 · dataset
Exact listed model name; hosted reasoning/quantization may differ from local checkpoint.
Scores describe the evaluated hosted variant, not each quantized file. AA and Arena are never averaged.
Reported speeds 1
| Hardware | Quant | Backend | Prompt t/s | Gen t/s | Date | Source |
|---|---|---|---|---|---|---|
| 4x NVIDIA A100-SXM4-40GB + 2 TB system RAM (multi-GPU + CPU expert offload with --cpu-moe) multi-device | UD-IQ1_S | llama.cpp CUDA | 12.9 | 0.1 | 2026-08 | computingforgeeks.com |
Reported by the linked sources, not reproduced by LocalFit. All rows in the speed database