Qwen 3.5 2B

Phone-class multimodal model built on the same Qwen 3.5 foundation as the medium tier. Non-thinking by default. Runs anywhere, including CPU-only setups.

License: Apache 2.0 · Context: 262K native · Released: March 2, 2026

The decision in five lines

The call: Skip for local — for coding
Best for: coding
Runs on: 23 hardware picks fit (cheapest: Intel Arc B580 12 GB · $249)
Watch out: Anything requiring multi-step reasoning or code beyond snippets — at 2B it hallucinates structure on anything non-trivial.
Evidence: Estimated · last verified April 2026

2B dense: PARAMETERS
DENSE: TYPE
262K: CONTEXT
~1.5 GB: VRAM AT Q4

Where we recommend this

Every tier slot in the planner where this model is a top or alternate pick. Pulled live from planner.js — when the planner refreshes, this table stays current.

CODING · LOW

Qwen 3.5 2BSmallest coherent option; 262K context; fine for autocomplete loops.

The call

Phone-class multimodal model built on the same Qwen 3.5 foundation as the medium tier. Non-thinking by default. Runs anywhere, including CPU-only setups.
When not to use: Anything requiring multi-step reasoning or code beyond snippets — at 2B it hallucinates structure on anything non-trivial. Utility model, not a thinker.

Runner notes

Ollama tag `qwen3.5:2b` (Q4_K_M default). Runs on phones, Raspberry Pi 5, and anything else with 4 GB+ RAM.

License: Apache 2.0
Released: March 2, 2026
Maker: Alibaba
Model card: huggingface.co/Qwen/Qwen3.5-2B →

Hardware that fits

Every hardware pick whose memory fits this model at the quant we recommend. Sorted cheapest-first — the top row is your best-value fit. Click through for the full buyer’s guide.

Next step

Find-by-model — see what hardware runs this→