Machine

Vega Series Vega Mobile Series · Ryzen 5 3400G #b8495f13

48 runs of 16 models since Oct 12, 2026. Last run Nov 13, 2026.

ShareRedditX
48
runs
16
models
46
configs
2
configs run 2+ times

Hardware

GPU
AMD Radeon Vega Series / Radeon Vega Mobile Series 2 GB
CPU
AMD Ryzen 5 3400G with Radeon Vega Graphics 4 cores
RAM
13 GB
OS
Ubuntu 26.04.1 LTS
Tools
Bionic 35 · Unsloth Studio 13
Client
0.4.99
All AMD Radeon Vega Series / Radeon Vega Mobile Series 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 read103050709051020tok/s (log scale) →quality →outliers sit at the edges

Medians of each model's best config, 16 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.

google/gemma-4-e2b
4 runs · 4 configs
★ bestBionic64k ctxthinking default
68.0
1 rated run
20.1 tok/s
ministral-3-8b-instruct-2512
2 runs · 2 configs
★ bestBionic8k ctxthinking default
67.6
1 rated run
6.6 tok/s
qwen3.8-4b-distill
4 runs · 4 configs
★ bestBionic16k ctxthinking default
63.1
1 rated run
6.4 tok/s
ministral-3-3b-instruct-2512
4 runs · 4 configs
★ bestBionic64k ctxthinking default
59.7
1 rated run
13.1 tok/s
gemma-3-4b-it
5 runs · 4 configs
★ bestBionic8k ctxthinking default
×2 runs
52.851.0–54.7
steady
12.7 tok/s
12.4–12.9
mistralai/ministral-3-3b
4 runs · 4 configs
★ bestBionic8k ctxthinking on
47.9
1 rated run
12.9 tok/s
qwen3.8-2b-distill
4 runs · 4 configs
★ bestBionic8k ctxthinking default
43.7
1 rated run
14.6 tok/s
unsloth/gemma-3-1b-it-GGUF
5 runs · 5 configs
★ bestQ8_0Unsloth Studio8k ctxthinking default
40.6
1 rated run
31.5 tok/s
1bit-MONSTER/SmolLM3-3B-GGUF
2 runs · 2 configs
★ bestQ4_K_MUnsloth Studio64k ctxthinking default
40.6
1 rated run
15.0 tok/s
prism-ml/Ternary-Bonsai-2-2B-gguf
1 run · 1 config
★ bestQ2_0Unsloth Studio16k ctxthinking default
36.9
1 rated run
35.3 tok/s

Compare with another machine

Run history

Show all 48 runs ▾