Fastest
qwen36-35b-ik
35B
152.1 tok/s
Highest token generation speed at 152.1 tok/s.
Hardware Benchmarks
Best Local LLM Performance for RTX 3090
Compare RTX 3090 benchmark results and see which models actually earn the best completed quality scores.
Hardware Overview
GPU-specific specifications for local LLM planning.
- VRAM
- 24GB
Recommended Models
Models we recommend for RTX 3090.
Best quality & Largest context
Qwen3.8-27B
27B
38.4 tok/sQuality 90.9Context 160,000 tokens
Highest completed quality score at 90.9. Largest context that fits in 24GB VRAM, at 160,000 tokens.
Frequently asked questions
Which models actually run best on the RTX 3090, by task, from community benchmark data.
- What is the best local model for coding on the RTX 3090?
- On the RTX 3090, Qwen3.8-27B-UD-IQ4_XS ranks first for coding in our benchmarks (79.4/100 at ~65.0 tok/s).
- What is the best local model for agentic workflows on the RTX 3090?
- On the RTX 3090, qwen3.8-flash-next-iq3_s ranks first for agentic workflows in our benchmarks (92.5/100 at ~79.6 tok/s). unsloth/Qwen3.6-35B-A3B-MTP-GGUF is faster (~117.4 tok/s) and still scores well (88.2/100) — a good pick if you'd rather trade a little quality for speed.
Benchmark Results
Completed quality-ranked benchmark results for NVIDIA GeForce RTX 3090.