VOOZH about

URL: https://www.hardware-corner.net/gpu-llm-benchmarks/rtx-3090/

⇱ RTX 3090 Local LLM Benchmarks, Context Scaling & Supported Models 2026 – Hardware Corner

Tier 1 Enthusiast

RTX 3090 LLM Performance

Local LLM Performance: 52.1 t/s average on 14B models at 16k context. Updated Benchmarks: March 2026.

Gen (14B 4-bit) 52.1 t/s
PP (14B 4-bit) 1,679 t/s
Max Model 34B
VRAM
24 GB GDDR6
Bandwidth 986 GB/s
Token Gen (14B @ 4k Ctx)

52.1T/s

Prompt Proc (14B @ 4k Ctx)

1,679T/s

Summary

Even in 2026, we find the NVIDIA RTX 3090 remains a powerhouse for local LLM workloads. In our testing, its 24GB of VRAM and high memory bandwidth allowed us to comfortably run 35B parameter models using 4-bit quantization (Q4). Based on our market analysis, we consider it the most cost-effective second-hand entry point for anyone serious about local LLM work.

Key Insights

Capable of running Qwen3 34B using GGUF (Q4_K) completely in VRAM.
Handles context splitting effectively up to 16k tokens on 34B model.
Handles gpt-oss 120B model in MXFP4 quantization at full128K context.
Supports Flash Attention 2, significantly boosting prompt processing speeds.

Current Price in US

$1,000

Avg. Market Value

Current Pricing

Hardware Specs
VRAM 24GB GDDR6
Capable of running 34B model
Bandwidth 986 GB/s
Architecture Ampere
Memory speed 19.6 Gbps
Memory bus 384 bit
TDP 350 W
Suggested PSU 750 W
Price/GB VRAM $41.67
Price/(t/s) with 14B @ 16k $19.18

Biggest LLMs You Can Run on This GPU

The models below represent the largest language models that fit fully in VRAM on this GPU using 4-bit quantization (GGUF). Benchmarks include token generation and prompt processing speeds measured at their maximum supported context length.

Qwen3.5 35B (MXFP4) Max 128k
Token Generation 79.4 t/s @ 128k context
Prompt Processing 1,288.9 t/s @ 128k context
Qwen3 32B (Q4_K) Max 16k
Token Generation 30.3 t/s @ 16k context
Prompt Processing 767.8 t/s @ 16k context
Gemma4 31B (Q4_K) Max 32k
Token Generation 31.4 t/s @ 32k context
Prompt Processing 723.7 t/s @ 32k context

Note: Context values are grouped into standard tiers (4K, 16K, 32K, 64K, 128K). Models may support slightly higher context, but they remain in the lower tier unless they reach the next bracket.

RTX 3090 local LLM Inference Performance vs Similar GPUs

Compare prompt ingestion and token generation speeds against similar GPUs across widely used local models and extended context lengths up to 256K.

Local LLM Benchmarks

Prompt processing (t/s) and token generation speed (t/s) across different open weight models and context lengths.

Prompt Processing
Model 4k Ctx 16k Ctx 32k Ctx 64k Ctx 128k Ctx 256k Ctx
Qwen3 8B (Q4_K)
CUDA, -fa 1
 4,049.6 2,572.5 1,714.6 1,014.3 570.0 —
Qwen3 14B (Q4_K)
CUDA, -fa 1
 2,459.0 1,678.7 1,175.7 734.1 — —
gpt-oss 20B (MXFP4)
CUDA, -fa 1
 4,400.3 3,243.6 2,547.2 1,720.6 923.8 —
Gemma4 26B (Q4_K) 3,625.6 3,068.9 2,453.4 1,765.1 1,147.1 671.4
Qwen3.5 27B (Q4_K) 1,104.2 977.4 848.2 678.9 — —
Qwen3 30B A3B (Q4_K)
CUDA, -fa 1
 2,988.6 1,959.0 1,336.8 800.9 — —
Gemma4 31B (Q4_K) 1,155.8 913.2 723.7 — — —
Qwen3 32B (Q4_K)
CUDA, -fa 1
 1,087.9 767.8 — — — —
Qwen3.5 35B (MXFP4) 2,622.1 2,381.3 2,121.6 1,749.8 1,288.9 —
Token Generation
Model 4k Ctx 16k Ctx 32k Ctx 64k Ctx 128k Ctx 256k Ctx
Qwen3 8B (Q4_K)
CUDA, -fa 1
 115.3 87.5 67.9 46.6 28.1 —
Qwen3 14B (Q4_K)
CUDA, -fa 1
 70.0 52.1 38.6 25.4 — —
gpt-oss 20B (MXFP4)
CUDA, -fa 1
 147.5 128.5 112.6 89.6 62.2 —
Gemma4 26B (Q4_K) 119.4 115.0 107.5 98.9 83.0 64.4
Qwen3.5 27B (Q4_K) 33.5 32.3 31.0 28.8 — —
Qwen3 30B A3B (Q4_K)
CUDA, -fa 1
 153.6 113.8 87.2 58.3 — —
Gemma4 31B (Q4_K) 34.7 33.5 31.4 — — —
Qwen3 32B (Q4_K)
CUDA, -fa 1
 35.1 30.3 — — — —
Qwen3.5 35B (MXFP4) 111.2 107.1 101.2 93.1 79.4 —

Frequently Asked Questions

Common questions about running LLMs on the RTX 3090.