Machine

2× RTX 5060 · Core i7-8700 #3a7b23fa

199 runs of 100 models since Sep 8, 2026. Last run Nov 21, 2026.

ShareRedditX
199
runs
100
models
118
configs
41
configs run 2+ times

Hardware

GPU
2× NVIDIA GeForce RTX 5060 16 GB
CPU
Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz 6 cores
RAM
31 GB
OS
Ubuntu 26.04.1 LTS
Tools
llama.cpp 191 · capy-flash 5 · PolyStrata 3
Client
0.4.99
All 2× NVIDIA GeForce RTX 5060 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, 100 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-EfficientThink-VAL-APEX-I-NanoPlus-V3
1 run · 1 config
★ bestllama.cpp128k ctxthinking oneffort medium
88.3
1 rated run
38.2 tok/s
Qwen3.8-27B-GSQ-RCO-IQ3_S
1 run · 1 config
★ bestllama.cpp128k ctxthinking on
87.6
1 rated run
25.6 tok/s
Qwen3.8-27B-i1-IQ4_XS-GGUF-Smaller
1 run · 1 config
★ bestllama.cpp64k ctxthinking oneffort xhigh
86.5
1 rated run
40.1 tok/s
Dirk-Qwen3.8-27B-UD-IQ4_XS
4 runs · 1 config
★ bestllama.cpp64k ctxthinking on
×4 runs
85.982.6–87.8
steady
37.5 tok/s
36.6–38.2
ornith-1.5-35b-a3b-apex-i-quality
1 run · 1 config
★ bestPolyStrata128k ctxthinking on
85.9
1 rated run
93.2 tok/s
VeriLoop-E2-GSQ-RCO-IQ3_S-mtp
1 run · 1 config
★ bestllama.cpp96k ctxthinking on
85.8
1 rated run
39.7 tok/s
Qwen3.8-27B-GSQ-RCO-IQ3_S-recovered
1 run · 1 config
★ bestllama.cpp128k ctxthinking on
85.7
1 rated run
25.6 tok/s
Qwen3.8-27B-4.2BPW-16GB
1 run · 1 config
★ bestllama.cpp80k ctxthinking on
85.5
1 rated run
34.6 tok/s
Qwen3.8-27B-YMQ-M-TI
6 runs · 2 configs
★ bestllama.cpp96k ctxthinking on
×2 runs
85.483.5–87.2
steady
36.8 tok/s
36.8–36.9
Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1
1 run · 1 config
★ bestllama.cpp128k ctxthinking on
85.3
1 rated run
57.2 tok/s

Compare with another machine

Run history

Show all 199 runs ▾