What is the best local model for coding on the INTEL Arc Pro B390?

On the INTEL Arc Pro B390, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q4_K_M ranks first for coding in our benchmarks (86.0/100 at ~22.3 tok/s).

What is the best local model for agentic workflows on the INTEL Arc Pro B390?

On the INTEL Arc Pro B390, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q4_K_M ranks first for agentic workflows in our benchmarks (90.2/100 at ~22.3 tok/s). Qwen3.6-35B-A3B-UD-Q4_K_M is faster (~26.6 tok/s) and still scores well (83.6/100) — a good pick if you'd rather trade a little quality for speed.

Hardware Benchmarks

Best Local LLM Performance for INTEL Arc Pro B390

Compare INTEL Arc Pro B390 benchmark results and see which models actually earn the best completed quality scores.

ShareReddit X

Hardware Overview

GPU-specific specifications for local LLM planning.

Intel Arc Pro B390

Recommended Models

Models we recommend for INTEL Arc Pro B390.

Fastest

flint-Q4_K_M

40B

37.0 tok/s

Highest token generation speed at 37.0 tok/s.

Best quality

Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q4_K_M

35B

22.3 tok/sQuality 86.4

Highest completed quality score at 86.4.

Frequently asked questions

Which models actually run best on the INTEL Arc Pro B390, by task, from community benchmark data.

What is the best local model for coding on the INTEL Arc Pro B390?: On the INTEL Arc Pro B390, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q4_K_M ranks first for coding in our benchmarks (86.0/100 at ~22.3 tok/s).
What is the best local model for agentic workflows on the INTEL Arc Pro B390?: On the INTEL Arc Pro B390, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q4_K_M ranks first for agentic workflows in our benchmarks (90.2/100 at ~22.3 tok/s). Qwen3.6-35B-A3B-UD-Q4_K_M is faster (~26.6 tok/s) and still scores well (83.6/100) — a good pick if you'd rather trade a little quality for speed.

Benchmark Results

Completed quality-ranked benchmark results for Intel Arc Pro B390.

Open run page →