Machine

RTX A6000 · Core i9-14900K #294df0f1

48 runs of 27 models since Oct 20, 2026. Last run Nov 5, 2026.

ShareRedditX
48
runs
27
models
33
configs
11
configs run 2+ times

Hardware

GPU
NVIDIA RTX A6000 47 GB
CPU
Intel(R) Core(TM) i9-14900K 24 cores
RAM
63 GB
OS
Windows 11 רҵ�� 25H2
Tools
llama.cpp 48
Client
0.4.98
All NVIDIA RTX A6000 benchmarks

Models on this machine

One dot per model, at its best config. Up is better answers, right is faster. Time to finish counts every token the model wrote, thinking included, so a model that talks a lot ends up further left.

Best trade-off: quality 60+, and nothing else here is both faster and betterOther models · dot size = number of runs
below 20 tok/s: slower than you read4050607080901002050100tok/s (log scale) →quality →

Medians of each model's best config, 27 models.

Best config per model

sort

Runs with the same model, quant, tool, context, thinking and KV cache count as one config. Numbers are the median of its runs, and each dot is one run on the 0–100 quality scale. ★ marks the best config with at least 2 rated runs.

Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp
1 run · 1 config
★ bestllama.cpp256k ctxthinking oneffort xhigh
88.5
1 rated run
39.4 tok/s
LuffyTheFox-Swift-Qwen3.8-27B-Genesis-GGUF-NVFP4
1 run · 1 config
★ bestllama.cpp256k ctxthinking oneffort xhigh
88.5
1 rated run
36.6 tok/s
JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M
3 runs · 1 config
★ bestQ5_K_Mllama.cpp256k ctxthinking oneffort xhigh
×3 runs
87.785.1–90.5
steady1 fast run
30.6 tok/s
30.0–50.8
Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2
3 runs · 1 config
★ bestllama.cpp256k ctxthinking oneffort xhigh
×3 runs
87.183.8–88.7
steady
26.4 tok/s
26.3–26.9
Qwen3.8-27B-Omnimerge-v6-Q4_K_L
1 run · 1 config
★ bestQ4_Kllama.cpp256k ctxthinking oneffort xhigh
87.0
1 rated run
33.4 tok/s
Qwen3.8-27B-Omnimerge-v6-MTP-GGUF
2 runs · 2 configs
★ bestllama.cpp262k ctxthinking oneffort xhigh
86.8
1 rated run
56.6 tok/s
Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2-GGUF-Q6_K
1 run · 1 config
★ bestQ6_Kllama.cpp256k ctxthinking oneffort xhigh
86.8
1 rated run
42.6 tok/s
Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp
4 runs · 2 configs
★ bestQ5_K_Mllama.cpp256k ctxthinking oneffort xhigh
×2 runs
85.985.8–86.0
steady
30.0 tok/s
29.7–30.2
Qwen3.8-27B-pi-GGUF-Q4_K_M
1 run · 1 config
★ bestQ4_K_Mllama.cpp256k ctxthinking oneffort xhigh
85.9
1 rated run
33.9 tok/s
Qwen3.8-27B-Cold-Fusion-GAIN-V1.1-NM-DAU-NEO-MAX-MTP-GGUF-Q6_K
2 runs · 1 config
★ bestQ6_Kllama.cpp256k ctxthinking oneffort xhigh
×2 runs
85.783.0–88.4
steady
34.2 tok/s
24.7–43.7

Compare with another machine

Run history

Show all 48 runs ▾