MiniCPM4-8B
Ultra-efficient 8B for edge devices with 7x faster decoding.
About
ModelBest's efficiency-first 8B model designed for phones and edge hardware: sparse attention (InfLLM v2), speculative sampling, and their CPM.cu inference engine.
It matches Qwen3-8B on standard benchmarks while training on just 8T tokens, and decodes roughly 7x faster on a Jetson AGX Orin. Apache 2.0 licensed.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
MMLU 86.38, MATH500 95.60, HumanEval 95.73, AIME24 83.33 per the MiniCPM4 technical report; ~7x faster decoding than Qwen3-8B on Jetson AGX Orin.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Strong adoption in the efficient-inference community; CPM.cu engine widely used; ModelBest maintains the series with quantized and speculative variants; Apache 2.0 weights and technical report; training data not fully disclosed. Manual review: benchmark table and 7x Jetson speedup claim taken from the technical report, not independently reproduced. License: Apache 2.0.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Strong adoption in the efficient-inference community; CPM.cu engine widely used
- · ModelBest maintains the series with quantized and speculative variants
- · Apache 2.0 weights and technical report; training data not fully disclosed
- · manual review: benchmark table and 7x Jetson speedup claim taken from the technical report, not independently reproduced