Machine
RX 7900 XTX · Ryzen Threadripper 3970X #213b1ee3
26 runs of 11 models since Nov 3, 2026. Last run Nov 8, 2026.
26
runs
11
models
23
configs
3
configs run 2+ times
Hardware
- GPU
- AMD Radeon RX 7900 XTX 24 GB
- CPU
- AMD Ryzen Threadripper 3970X 32-Core Processor 32 cores
- RAM
- 94 GB
- OS
- Ubuntu 26.04.1 LTS
- Tools
- OpenAI-compatible 19 · Strata 7
- Client
- 0.4.99
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, 11 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
Qwen3.8-Flash-Next-Avalon-IQ3KS-00001-of-00033
1 run · 1 config
strata_iq3s_128k
1 run · 1 config
strata_iq3s_256k
2 runs · 2 configs
strata_ud-q5
1 run · 1 config
qwen3.8-flash-next-unsloth-ud-iq4_xs-v31
6 runs · 6 configs
default
4 runs · 4 configs
strata_swift
4 runs · 3 configs
★ bestOpenAI-compatible32k ctxthinking on
×2 runs
84.684.4–84.8
steady
56.3 tok/s
55.6–57.0
ThinkingCap-Qwen3.8-27B
2 runs · 2 configs
strata_iq4
2 runs · 1 config
★ bestOpenAI-compatible32k ctxthinking on
×2 runs
83.880.5–87.2
steady
60.1 tok/s
54.3–66.0
cyber-frost-3.8-ps
2 runs · 1 config
★ bestStrata64k ctxthinking defaulteffort xhigh
×2 runs
83.783.4–84.1
steady
27.0 tok/s
24.1–29.9
Compare with another machine
Run history
Nov 8, 2026
20:47cyber-frost-3.8-ps×2Strata27.0 tok/s83.709:29qwen3.8-flash-next-unsloth-ud-iq4_xs-v31Strata58.4 tok/s85.001:58qwen3.8-flash-next-unsloth-ud-iq4_xs-v31OpenAI-compatible66.2 tok/s79.4Nov 7, 2026
23:48qwen3.8-flash-next-unsloth-ud-iq4_xs-v31OpenAI-compatible59.6 tok/s72.900:15Qwen3.8-Flash-Next-Avalon-IQ3KS-00001-of-00033OpenAI-compatible61.7 tok/s88.3Nov 6, 2026
23:50qwen3.8-flash-next-unsloth-ud-iq4_xs-v31Strata64.6 tok/s86.323:18qwen3.8-flash-next-unsloth-ud-iq4_xs-v31Strata63.7 tok/s85.822:58qwen3.8-flash-next-unsloth-ud-iq4_xs-v31Strata60.5 tok/s83.222:36qwen3.8-flash-next-iq3_sStrata67.3 tok/s69.911:47strata_iq4×2OpenAI-compatible60.1 tok/s83.8Nov 5, 2026
12:45strata_iq3s_256kOpenAI-compatible39.8 tok/s86.812:18strata_iq3s_256kOpenAI-compatible40.2 tok/s86.111:40strata_swiftOpenAI-compatible56.2 tok/s83.311:23strata_swiftOpenAI-compatible56.8 tok/s85.111:00strata_swift×2OpenAI-compatible56.3 tok/s84.607:14ThinkingCap-Qwen3.8-27BOpenAI-compatible46.5 tok/s84.406:49strata_iq3s_128kOpenAI-compatible53.7 tok/s87.2