Models
Meta's ParetoQ compares five bit widths and puts the accuracy cliff at 1 bit
Meta trained the same models at 1, 1.58, 2, 3 and 4 bits under one recipe and reports 1.58-bit to 3-bit beating 4-bit at matched memory.
Amazon survey puts some small models at 10 to 100 times their parameter count
A survey of about 160 papers finds 1B to 8B models matching much larger ones on single benchmarks, and shows where models below 1B fall apart.
BlueLM-V-3B runs a multimodal model on a Dimensity 9300 NPU in 2.2 GB
CUHK and vivo AI Lab report 24.4 tok/s and 2.2 GB peak memory for a 3B vision-language model on a MediaTek Dimensity 9300.
PhoneLM searches for a fast architecture before training it and hits 58 tok/s
BUPT researchers picked their 0.5B and 1.5B transformer shapes by measuring speed on a Snapdragon 8 Gen 3 first, then pre-training the winner.
Hugging Face trains SmolLM2 at 135M, 360M and 1.7B on up to 11T tokens
Hugging Face released SmolLM2 in three sizes trained on up to 11 trillion tokens, with 4-bit builds from 118 MB for on-device runtimes.
Apple team finds H100 last on tokens per dollar for models up to 2B
Seven Apple authors measured tokens per dollar for LLaMA-style models from 100M to 2B and found H100s last at every size, behind cheaper A100s.
Mistral puts Ministral 3B and 8B on devices with 128k context
Ministral 3B and 8B handle up to 128k tokens for on-device work, but only the 8B Instruct weights were published, and for research use.
AMD trains its first small language model, AMD-Llama-135M, on MI250 accelerators
AMD trained a 135M model from scratch on Instinct MI250 accelerators and reports up to 3.88x faster CodeLlama-7b inference when it drafts tokens.
Meta releases Llama 3.2 1B and 3B for phones and edge devices
The two lightweight models carry a 128K context window, were pruned and distilled from Llama 3.1, and shipped with day-one Qualcomm and MediaTek support.
Ai2 releases OLMoE, 7B parameters with 1B active per token
Ai2's mixture-of-experts model holds 6.9B parameters but runs 1.3B per token, and its iOS app runs it offline on an iPhone 15 Pro or newer.
Gemma 2 2B is distilled from a larger model and scores 1126 on Chatbot Arena
Google released a 2.6B-parameter Gemma 2 trained by distilling a larger model, and reported an Elo of 1126 on the LMSYS Chatbot Arena.
Spectra ships ternary 3.9B models that match 4-bit quantisation at half the bits
Nolano AI trained 54 models from 99M to 3.9B parameters on the same 300B tokens in ternary, quantised and half-precision form to compare them by bit size.