Machine

680m · Ryzen 5 6600H #b420cc10

21 runs of 14 models since Oct 30, 2026. Last run Nov 14, 2026.

ShareRedditX
21
runs
14
models
14
configs
4
configs run 2+ times

Hardware

GPU
AMD Radeon 680M 4 GB
CPU
AMD Ryzen 5 6600H with Radeon Graphics 6 cores
RAM
11 GB
OS
Arch Linux
Tools
llama.cpp 21
Client
0.4.99
All AMD Radeon 680M 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 read40506070809010051020tok/s (log scale) →quality →

Medians of each model's best config, 14 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.

Qwen3.6-35B-A3B-UD-Q2_K_XL
1 run · 1 config
★ bestQ2_Kllama.cpp8k ctxthinking off
76.6
1 rated run
13.4 tok/s
Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q2_K_XL
2 runs · 1 config
★ bestQ2_Kllama.cpp8k ctxthinking off
×2 runs
75.469.4–81.4
varies
14.0 tok/s
14.0–14.1
Cyber-Tiel-Coder-35B-A3B-MTP-IQ3_S
1 run · 1 config
★ bestllama.cpp8k ctxthinking off
73.9
1 rated run
11.2 tok/s
cyber-tiel-coder-q2kxl
1 run · 1 config
★ bestllama.cpp8k ctxthinking default
72.1
1 rated run
12.9 tok/s
Cyber-Tiel-Coder-35B-A3B-UD-IQ3_S
5 runs · 1 config
★ bestllama.cpp8k ctxthinking off
×5 runs
71.766.7–79.3
varies
11.2 tok/s
10.7–11.6
Qwen3.6-35B-A3B-UD-IQ3_S
2 runs · 1 config
★ bestllama.cpp8k ctxthinking off
×2 runs
71.468.2–74.7
steady
11.2 tok/s
11.2–11.2
ds-v4-mtp-q4km
2 runs · 1 config
★ bestllama.cpp8k ctxthinking default
×2 runs
71.067.9–74.2
steady
8.9 tok/s
8.8–8.9
Qwen3.5-9B-Defiant-Fable-MTP-Q4_K_M
1 run · 1 config
★ bestQ4_K_Mllama.cpp8k ctxthinking off
68.1
1 rated run
6.0 tok/s
Qwen3.6-35B-A3B-UD-IQ3_XXS
1 run · 1 config
★ bestllama.cpp8k ctxthinking off
65.7
1 rated run
12.0 tok/s
Underdog-Saluki-27B-1.0-IQ2-mix
1 run · 1 config
★ bestllama.cpp8k ctxthinking off
65.3
1 rated run
3.5 tok/s

Compare with another machine

Run history