will it local

Best local LLMs for the GeForce MX150

⚠ Laptop part — mobile clocks and thermals run below desktop sibling speeds; treat the band as optimistic for sustained load.

Speed

computed band 2.6-4.7 tok/s for 8B Q4_K_M (roofline, 48 GB/s VRAM)

Effective decode window: 0.28–0.50 of 48 GB/s nominal → ~13–24 GB/s effective (llama.cpp decode, Q4_K_M basis; arch window, community-calibrated)

What fits (computed)

→ 13b-q4 no 14.6 GB vs 2 GB usable — does not fit

→ 3b-q4 no 2.8 GB vs 2 GB usable — does not fit

→ 4b-q4 no 3.7 GB vs 2 GB usable — does not fit

→ 8b-q4 no 6.1 GB vs 2 GB usable — does not fit

Cited community benches

no single-card bench published — band is computed (see notes)

Runs fully in memory (machine alternatives)

Nominal bandwidth: 48 GB/s — see the effective decode window above

→ qwen3:0.6b on Tesla M60 (16GB, used) ($35) 0.6B · 0.6GB · needs ~1GB (4k ctx)

→ qwen3:1.7b on Tesla M60 (16GB, used) ($35) 1.7B · 1.4GB · needs ~1.8GB (4k ctx)

→ llama3.2:1b on Tesla M60 (16GB, used) ($35) 1B · 1.3GB · needs ~1.4GB (4k ctx)

→ gemma3:1b on Tesla M60 (16GB, used) ($35) 1B · 1.0GB · needs ~1.1GB (4k ctx)

Borderline — runs, but offloads

→ llama3.2:3b 3B — 2.4 GB vs 2 GB usable — partial CPU offload, expect large speed loss

→ whisper:tiny tiny — 2.1 of 2 GB usable — barely over; real with q8 KV cache (halves KV) or shorter context

→ whisper base — 2.3 GB vs 2 GB usable — partial CPU offload, expect large speed loss

→ kokoro:82m 82M — 2.3 GB vs 2 GB usable — partial CPU offload, expect large speed loss

Gotchas

⚠ 2 GB VRAM is a hard wall — nothing beyond tiny 1B-class quants even partially offloaded

⚠ Underclocked '1D12' MX150 variant was silently shipped in some ultrabooks (lower clocks than the N17S-G1)

⚠ Rare 4 GB GDDR5 variants exist (Notebookcheck lists max 4 GB) but the vast majority are 2 GB — verify per laptop

⚠ Pascal: no tensor cores; pre-Turing CUDA is deprecated by many current LLM stacks — treat as display/CUDA-toy GPU

Where to go next

—