Can MacBook Air M4 32 GB run Recursive Flint (40B-A2.5B)?

Q: Can MacBook Air M4 32 GB run Recursive Flint (40B-A2.5B)?

Yes, excellent for chat; borderline for coding questions. Rated by Local AI Advisor's engine at MXFP4, about 40–55 tok/s for chat.

32 GB unified · 120 GB/s memory bandwidth · Recursive Flint (40B-A2.5B) has 39.8B parameters (2.5B active per token).

Short answer

Yes, excellent for chat; borderline for coding questions.

Best quantization for chat: MXFP4 (20 GB of weights) on LM Studio. Generation: 40–55 tok/s, leaving about 3.5 GB free for your other apps.

How it rates for each workload

Same model and machine, different jobs. Agents and long documents send far bigger prompts than chat, so they need much more speed and memory.

Workload	Verdict	Quant	Generation	First reply	Context
Chatwith Open WebUI	Excellent	MXFP4	40–55 tok/s	≈2 s	8K	Details
Coding questionswith Continue	Borderline	MXFP4	35–50 tok/s	≈8 s	16K	Details
Coding in a repositorywith Aider	Borderline	MXFP4	30–40 tok/s	≈25 s	32K	Details
Agentic codingwith OpenCode	Borderline	MXFP4	30–40 tok/s	≈35 s	32K	Details
Long documents (64K)with Open WebUI	Borderline	MXFP4	20–25 tok/s	≈2 min	64K	Details

Speeds are estimates calibrated against measured benchmarks on similar hardware. “First reply” is the time to read the workload’s opening prompt. How we calculate

Other models for MacBook Air M4 32 GB

The best-rated models for chat on this machine.

Everything MacBook Air M4 32 GB can run

Cheapest machines for Recursive Flint (40B-A2.5B)

Hardware rated comfortable or better for chat, lowest price first.

Same chip, other memory: 16 GB, 24 GB

Different tool, context size or workload?

These pages use typical settings. The full check models your exact setup.

Open sourceCheck my exact setup