Perceptron Mk1 alternatives
Quick answer · Updated July 2026
for a cheaper option, DeepSeek V4 Flash runs $0.090/M vs $0.150/M; for more speed, Morph V3 Large hits 3.6k t/s; and the best open-weight substitute is GLM 5.2.
Cheaper alternatives
deepseek-v4-flash
$0.090/M input · 40.3 intel
mimo-v2.5
$0.105/M input · 40.1 intel
minimax-m2.5
$0.120/M input · 33.7 intel
qwen3.6-35b-a3b
$0.140/M input · 31.6 intel
ring-2.6-1t
$0.075/M input · 30.6 intel
qwen3.5-35b-a3b
$0.140/M input · 29.3 intel
Faster alternatives
Open-source alternatives
glm-5.2
51.1 intel · open weights · 1.0M ctx
minimax-m3
44.4 intel · open weights · 1.0M ctx
deepseek-v4-pro
44.3 intel · open weights · 1.0M ctx
kimi-k2.6
42.8 intel · open weights · 262K ctx
mimo-v2.5-pro
42.2 intel · open weights · 1.0M ctx
kimi-k2.7-code
41.9 intel · open weights · 262K ctx
Frequently asked
Is there a cheaper alternative to Perceptron Mk1?
Yes — DeepSeek V4 Flash costs $0.090 per million input tokens versus $0.150 for Perceptron Mk1, while staying within ~8 points of its Intelligence Index.
Is there an open-source alternative to Perceptron Mk1?
GLM 5.2 is the strongest open-weight alternative to Perceptron Mk1 — downloadable from Hugging Face and self-hostable, scoring 51.1 on intelligence.