What can you run with 8GB of RAM?

With 8GB RAM you can run 51 of the 129 open-source models we track — up to about 16B parameters at Q4 quantization. The most popular that fits is Qwen3.5-4B, needing 3.8 GB and generating around ~12 tok/s on typical hardware of this size.

Typical machines: Base MacBook Air (M1/M2) · Budget Windows laptops · Entry-level desktops. After the operating system takes its share, about 6.4 GB is free for a model.

Having the memory is not the same as having the speed

8GB RAM decides whether a model loads. How fast it answers is decided by memory bandwidth — how quickly your machine can read the model, which it must do once for every word it writes. Machines with 8GB RAM span 51–102 GB/s, so the fastest is about 2.0× quicker than the slowest with identical capacity. The table quotes a DDR4-3200; pick yours below to see the difference. Why bandwidth and not the processor →

51 GB/sDesktop or older laptop — DDR4-3200· 3,200 MT/s × 128-bit dual channel ÷ 8 = 51.2 GB/s
68 GB/sThin laptop — LPDDR4X-4266· 4,266 MT/s × 128-bit ÷ 8 = 68.3 GB/s (soldered, 2 channels)
68 GB/sMac — M1· Apple published specification for the M1: 68.25 GB/s
90 GB/sDesktop or laptop — DDR5-5600· 5,600 MT/s × 128-bit dual channel ÷ 8 = 89.6 GB/s
100 GB/sMac — M2· Apple published specification for the M2: 100 GB/s
102 GB/sModern thin laptop — LPDDR5-6400· 6,400 MT/s × 128-bit ÷ 8 = 102.4 GB/s (soldered, 2 channels)

Models that run on 8GB RAM in 2026

ModelSizeMemoryEst. speedDownloadsFeels like
granite-embedding-97m-multilingual-r2Ibm97M603 MBRuns well~563 tok/s97,079Fast
LFM2.5-Encoder-230MLiquid Ai230M793 MBRuns well~239 tok/s13,432Fast
LFM2.5-230MLiquid Ai230M808 MBRuns well~216 tok/s55,999Fast
granite-embedding-311m-multilingual-r2Ibm312M821 MBRuns well~131 tok/s64,530Fast
granite-speech-5.0-470m-turboctcIbm473M857 MBRuns well~116 tok/s11,294Fast
harrier-oss-v1-270mMicrosoft268M865 MBRuns well~131 tok/s727,222Fast
LFM2.5-Encoder-350M-Policy-LinterLiquid Ai355M885 MBRuns well~155 tok/s9,957Fast
LFM2.5-VL-450MLiquid Ai449M900 MBRuns well~145 tok/s67,396Fast
nemotron-3.5-asr-streaming-0.6bNVIDIA638M1.1 GBRuns well~67 tok/s917,259Fast
Qwen3.5-0.8BAlibaba / Qwen873M1.3 GBRuns well~62 tok/s2,881,964Fast
privacy-filterOpenAI · MoE1.4B1.4 GBRuns well~196 tok/s280M active456,933Fast
LFM2.5-1.2B-InstructLiquid Ai1.2B1.4 GBRuns well~45 tok/s366,847Fast
LFM2.5-VL-1.6BLiquid Ai1.6B1.4 GBRuns well~45 tok/s248,110Fast
Nemotron-3-Embed-1B-BF16NVIDIA1.1B1.5 GBRuns well~48 tok/s386,914Fast
Hy-MT2-1.8BTencent2B1.9 GBRuns well~29 tok/s24,475Fast
MiniMax-Music3MiniMax2.4B2.1 GBRuns well~24 tok/s23,217Comfortable
Qwen3.5-2BAlibaba / Qwen2.3B2.1 GBRuns well~24 tok/s3,097,484Comfortable
cohere-transcribe-arabic-07-2026Cohere2.1B2.3 GBRuns well~21 tok/s51,893Comfortable
cohere-transcribe-03-2026Cohere2.1B2.3 GBRuns well~21 tok/s607,964Comfortable
granite-speech-4.1-2bIbm2.3B2.4 GBRuns well~21 tok/s237,181Comfortable
LFM2.5-2.6BLiquid Ai2.7B2.5 GBRuns well~20 tok/s133,677Comfortable
LFM2.5-VL-3BLiquid Ai3.1B2.5 GBRuns well~20 tok/s24,282Comfortable
granite-4.0-1b-speechIbm2.3B2.5 GBRuns well~21 tok/s74,644Comfortable
LFM2.5-Audio-1.5BLiquid Ai1.5B2.6 GBRuns well~18 tok/s1,702Comfortable
Unlimited-OCRBaidu AI3.3B2.7 GBRuns well~17 tok/s3,008,635Comfortable
Nemotron-Labs-Diffusion-3BNVIDIA3B2.8 GBRuns well~18 tok/s35,414Comfortable
LocateAnything-3BNVIDIA3.8B2.8 GBRuns well~16 tok/s95,054Comfortable
granite-4.1-3bIbm3.4B2.9 GBRuns well~16 tok/s122,076Comfortable
granite-4.2-3bIbm3.7B3.1 GBRuns well~15 tok/s14,073Comfortable
granite-4.0-3b-visionIbm4B3.1 GBRuns well~15 tok/s1,069Comfortable
Shieldstral-1.0-3BMistral AI3.8B3.1 GBRuns well~15 tok/s20,895Comfortable
Ministral-3-3B-Instruct-2512Mistral AI3.8B3.1 GBRuns well~15 tok/s544,419Comfortable
Nemotron-Labs-Diffusion-3B-BaseNVIDIA3.8B3.3 GBRuns well~14 tok/s19,213Comfortable
Cosmos3-EdgeNVIDIA3.9B3.3 GBRuns well~14 tok/s1,325,763Comfortable
Nemotron-3.5-Content-SafetyNVIDIA4.3B3.6 GBRuns well~13 tok/s134,547Comfortable
gemma-4-E2B-itGoogle DeepMind5.1B3.8 GBRuns well~11 tok/s3,235,641Comfortable
Qwen3.5-4BAlibaba / Qwen4.7B3.8 GBRuns well~12 tok/s7,681,585Comfortable
Fara1.5-4BMicrosoft4.5B4.0 GBRuns well~11 tok/s2,786Comfortable
MolmoWeb-4BAi24.9B4.1 GBRuns well~11 tok/s2,152Comfortable
NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA4B4.1 GBRuns well~12 tok/s712,701Comfortable
Fara-7BMicrosoft8.3B5.5 GBRuns well~7.1 tok/s1,593Slow
Hy-MT2-7BTencent8B5.7 GBTight~7.2 tok/s12,004Slow
LFM2.5-8B-A1BLiquid Ai · MoE8.5B5.9 GBTight~54 tok/s1B active96,675Fast
Nemotron-3-Embed-8B-BF16NVIDIA8B5.9 GBTight~6.9 tok/s81,804Slow
Nemotron-Labs-Diffusion-8BNVIDIA8B5.9 GBTight~6.9 tok/s163,233Slow
Cosmos3-NanoNVIDIA16B6.2 GBTight~6.6 tok/s244,209Slow
gemma-4-E4B-itGoogle DeepMind8B6.2 GBTight~6.2 tok/s4,740,694Slow
Nemotron-Labs-Diffusion-8B-BaseNVIDIA8.5B6.2 GBTight~6.5 tok/s280,417Slow
Ministral-3-8B-Instruct-2512Mistral AI8.9B6.3 GBTight~6.4 tok/s119,379Slow
granite-4.1-8bIbm8.8B6.3 GBTight~6.5 tok/s1,441,995Slow
MolmoWeb-8BAi28.7B6.4 GBTight~6.3 tok/s1,679Slow

Sorted by memory, smallest first — click any column to change it.

Upgrade

Including Qwen3.5-9B (9.7B), gemma-4-12B-it (12B) and Ornith-1.0-9B (9B) — around ~10 tok/s on a typical machine that size.

See everything 16GB RAM runs →

All figures at Q4_K_M quantization, 4K context. “Memory” includes the model, its context cache and runtime overhead. Speed is calculated, not benchmarked: 51 GB/s × 0.65 ÷ bytes read per token, assuming a DDR4-3200. Scale it by your own machine's bandwidth from the list above. Mixture-of-experts models read only their active parameters per token, which is why some large ones outrun smaller dense models.

Image generators that also fit

These make pictures rather than text. Their memory use scales with the resolution you generate at, not conversation length, so treat these figures as a floor rather than a ceiling.

What you'd need more memory for

The next models up, and what they ask for.

granite-4.2-8bneeds 6.6 GB
Ornith-1.0-9Bneeds 6.7 GB
Qwen3.5-9Bneeds 6.8 GB
GLM-4.6V-Flashneeds 7.0 GB
Fara1.5-9Bneeds 7.0 GB

Other hardware

Verified against HuggingFace on 2026-09-04New to this? Start here →Runs on a laptop and up · last 12 months only