Best Local LLMs for 8GB RAM

Ranked open-weight models that run well on a 8GB machine, with estimated speed and the run command for each.

Try: Qwen3 8B, RTX 4090, MacBook Pro M4, 16GB, iPhone 17 Pro

Ranked picks

Best local models for ~8GB

12 picks

Estimates assume a representative Apple-Silicon machine with 8GB unified memory. Tok/s are ModelFit estimates, not measured benchmarks; typical error is the median gap to measured runs on hardware the model was not fitted on (how the estimates work). Run the wizard for figures tuned to your exact chip.

  1. 01

    Qwen3.5 4B Instruct

    Qwen / 4B / Q4_K_M / ~3.5 GB

    Best for coding, agents, multimodal. Strong fit for 8 GB RAM with balanced speed and quality.

    ollama
    $ollama run qwen3.5:4b
    Fit
    Runs wellB · 37
    Best for
    Coding, Agents, Multimodal
    Perf
    est. 30 tok/s, typical error 10% · first token ~2.7s
  2. 02

    Gemma 4 E4B

    Gemma / 4.5B / Q4_K_M / ~4 GB

    Best for on-device, mobile, chat. Strong fit for 8 GB RAM with balanced speed and quality.

    ollama
    $ollama run gemma4:e4b
    Fit
    Runs wellB · 29
    Best for
    On-device, Mobile, Chat
    Perf
    est. 24 tok/s, typical error 10% · first token ~5.2s
  3. 03

    LFM2.5 8B-A1B

    LFM2 / 8.3B / Q4_K_M / ~5.5 GB

    This model may feel memory-heavy on 8 GB RAM, but it is still listed for balanced speed and quality.

    ollama
    $ollama run lfm2.5:8b-a1b-q4_K_M
    Fit
    Runs wellC · 2
    Best for
    On-device agents, tool calling, multilingual chat
    Perf
    est. 49 tok/s, typical error 10% · first token ~1.1s
  4. 04

    Qwen3.5 4B Instruct (Q8)

    Qwen / 4B / Q8_0 / ~4.3 GB

    This model may feel memory-heavy on 8 GB RAM, but it is still listed for balanced speed and quality.

    ollama
    $ollama run qwen3.5:4b-q8_0
    Fit
    Runs wellC · 23
    Best for
    Coding, Agents, Multimodal
    Perf
    est. 18 tok/s, typical error 10% · first token ~2.8s
  5. 05

    Gemma 4 E2B (Q8)

    Gemma / 2.3B / Q8_0 / ~4.6 GB

    This model may feel memory-heavy on 8 GB RAM, but it is still listed for balanced speed and quality.

    ollama
    $ollama run gemma4:e2b-it-q8_0
    Fit
    Runs wellC · 18
    Best for
    IoT, Mobile, Edge
    Perf
    est. 40 tok/s, typical error 10% · first token ~2.9s
  6. 06

    Ministral 3 3B

    Mistral / 3.4B / Q4_K_M / ~3 GB

    Best for chat, vision, edge. Strong fit for 8 GB RAM with balanced speed and quality.

    ollama
    $ollama run ministral-3:3b
    Fit
    Runs wellB · 46
    Best for
    Chat, Vision, Edge
    Perf
    est. 25 tok/s, typical error 10% · first token ~2.7s
  7. 07

    Gemma 4 E2B

    Gemma / 2.3B / Q4_K_M / ~2.3 GB

    Best for iot, mobile, edge. Strong fit for 8 GB RAM with balanced speed and quality.

    ollama
    $ollama run gemma4:e2b
    Fit
    Runs wellB · 59
    Best for
    IoT, Mobile, Edge
    Perf
    est. 55 tok/s, typical error 10% · first token ~2.8s
  8. 08

    Granite 4.2 8B

    Granite / 8.8B / Q4_K_M / ~5 GB

    This model may feel memory-heavy on 8 GB RAM, but it is still listed for balanced speed and quality.

    ollama
    $ollama run granite4.2:8b
    Fit
    Runs wellC · 11
    Best for
    Chat, Coding
    Perf
    est. 15 tok/s, typical error 10% · first token ~6.6s
  9. 09

    Qwen3.5 2B Instruct

    Qwen / 2B / Q4_K_M / ~1.8 GB

    Best for chat, edge tasks. Strong fit for 8 GB RAM with balanced speed and quality.

    ollama
    $ollama run qwen3.5:2b
    Fit
    Lots of roomA · 68
    Best for
    Chat, Edge tasks
    Perf
    est. 40 tok/s, typical error 5% · first token ~1.2s
  10. 10

    Ornith 1.0 9B

    Ornith / 9B / Q4_K_M / ~5.6 GB

    This model may feel memory-heavy on 8 GB RAM, but it is still listed for balanced speed and quality.

    ollama
    $ollama run ornith:9b
    Fit
    Runs wellC · 0
    Best for
    Agentic coding on small machines
    Perf
    est. 17 tok/s, typical error 10% · first token ~5.8s
  11. 11

    Granite 4.2 3B

    Granite / 3.7B / Q4_K_M / ~2.1 GB

    Best for chat, coding. Strong fit for 8 GB RAM with balanced speed and quality.

    ollama
    $ollama run granite4.2:3b
    Fit
    Lots of roomA · 63
    Best for
    Chat, Coding
    Perf
    est. 34 tok/s, typical error 10% · first token ~2.8s
  12. 12

    Phi-4 Mini 3.8B

    Phi / 3.8B / Q4_K_M / ~3.2 GB

    Best for coding, chat. Strong fit for 8 GB RAM with balanced speed and quality.

    ollama
    $ollama run phi4-mini:3.8b
    Fit
    Runs wellB · 43
    Best for
    Coding, Chat
    Perf
    est. 28 tok/s, typical error 10% · first token ~2.9s

Browse by RAM tier

Want an exact recommendation?

The wizard tunes picks and speed estimates to your exact device, chip, and RAM.

Open ModelFit Wizard Browse all devices
Best Local LLMs for 8GB RAM (2026) | ModelFit