Machine
RX 7900 XTX · Ryzen 7 7800X3D #05353333
18 runs of 10 models since Sep 27, 2026. Last run Oct 21, 2026.
18
runs
10
models
13
configs
4
configs run 2+ times
Hardware
- GPU
- AMD Radeon RX 7900 XTX 24 GB
- CPU
- AMD Ryzen 7 7800X3D 8-Core Processor 8 cores
- RAM
- 61 GB
- OS
- CachyOS
- Tools
- llama.cpp 11 · Strata 5 · OpenAI-compatible 2
- Client
- 0.4.96
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
Medians of each model's best config, 10 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.
configrunsqualityeach runspeed
unsloth/Qwen3.8-27B-GGUF:Qwen3.8-27B-UD-Q4_K_S
1 run · 1 config
byteshape/Qwen3.8-27B-GGUF:Qwen3.8-27B-IQ4_XS-3.84bpw
1 run · 1 config
swift-iq3_s
1 run · 1 config
ukisai/Swift-1.5-Qwen3.8-27B-GGUF:Swift-1.5-Qwen3.8-27B-Q4_K_S
3 runs · 2 configs
★ bestQ4_K_Sllama.cpp64k ctxthinking oneffort xhigh
×2 runs
86.185.4–86.9
steady
70.8 tok/s
69.2–72.5
iq3_s
1 run · 1 config
IsValorum/Ornith-1.5-35B-A3B-APEX-I-MiniPlus-V2.1-GGUF
1 run · 1 config
swift-1.5-iq3_xxs
5 runs · 3 configs
★ bestStrata64k ctxthinking oneffort medium
×3 runs
84.481.3–85.6
steady
53.0 tok/s
52.7–55.5
IsValorum/Nex-N2.5-mini-APEX-I-MiniPlus-V2.1-GGUF
2 runs · 1 config
★ bestllama.cpp64k ctxthinking on
×2 runs
84.283.6–84.8
steady
136 tok/s
135–136
nex-agi_Nex-N2.5-mini-IQ4_XS
1 run · 1 config
ukisai/Swift-1.5-Qwen3.8-27B-GGUF:Swift-1.5-Qwen3.8-27B-IQ4_XS
2 runs · 1 config
★ bestllama.cpp64k ctxthinking oneffort xhigh
×2 runs
83.579.2–87.9
varies
71.5 tok/s
71.1–71.9
Compare with another machine
Run history
Oct 21, 2026
10:21iq3_sOpenAI-compatible56.1 tok/s85.1Oct 20, 2026
13:42swift-iq3_sOpenAI-compatible56.3 tok/s86.4Oct 5, 2026
13:09Swift-1.5-Qwen3.8-27B-GGUF:Swift-1.5-Qwen3.8-27B-Q4_K_Sllama.cpp75.5 tok/s85.313:05Swift-1.5-Qwen3.8-27B-GGUF:Swift-1.5-Qwen3.8-27B-Q4_K_S×2llama.cpp70.8 tok/s86.1Oct 3, 2026
16:15swift-1.5-iq3_xxsStrata52.7 tok/s81.316:00swift-1.5-iq3_xxsStrata50.8 tok/s84.915:36swift-1.5-iq3_xxsStrata55.5 tok/s85.615:07swift-1.5-iq3_xxsStrata53.2 tok/s84.014:51swift-1.5-iq3_xxsStrata53.0 tok/s84.4Sep 27, 2026
08:17Swift-1.5-Qwen3.8-27B-GGUF:Swift-1.5-Qwen3.8-27B-IQ4_XS×2llama.cpp71.5 tok/s83.506:37Qwen3.8-27B-GGUF:Qwen3.8-27B-UD-Q4_K_Sllama.cpp67.7 tok/s89.106:06Qwen3.8-27B-GGUF:Qwen3.8-27B-IQ4_XS-3.84bpwllama.cpp64.0 tok/s87.905:40Ornith-1.5-35B-A3B-APEX-I-MiniPlus-V2.1-GGUFllama.cpp127 tok/s85.005:14Nex-N2.5-mini-APEX-I-MiniPlus-V2.1-GGUF×2llama.cpp136 tok/s84.204:42nex-agi_Nex-N2.5-mini-IQ4_XSllama.cpp126 tok/s83.6