llmrun

NVIDIA · Turing · 48 GB VRAM tier

Best AI Models for NVIDIA Quadro RTX 8000 (48 GB)

Runs 27–31B and 35B MoE models well; 70B barely fits and is slow (11 tok/s).

48 GB GDDR6672 GB/s65 TFLOPS FP16260 WTuring 2018

Cheapest single-card 48 GB with Tensor Cores, but Turing-era — no FP8/BF16 acceleration and slower than Ampere+ workstation cards.

What LLMs can the Quadro RTX 8000 run?

#ModelSpeedllmrun ScoreQuant · VRAM
1
Qwen3.8 27BAlibabaHow to run →
Q4_K_M · 17.9 GB · 25.1 t/s
25.1 tok/sQ4_K_M · 17.9 GB
2
Qwen3.6 27BAlibaba
Q4_K_M · 17.9 GB · 25.1 t/s
25.1 tok/sQ4_K_M · 17.9 GB
3
Gemma 4 31B ITGoogle
Q4_K_M · 21.7 GB · 21.4 t/s
21.4 tok/sQ4_K_M · 21.7 GB
4
Gemma 4 26B A4B ITGoogle
Q4_K_M · 16.5 GB · 126 t/s
126 tok/sQ4_K_M · 16.5 GB
5
NVIDIA Nemotron 3.5 Lightning 30B A3B BF16NVIDIA
Q4_K_M · 19.3 GB · 162 t/s
162 tok/sQ4_K_M · 19.3 GB
See all 368 models on the Quadro RTX 8000 →Best quant that fits · speeds estimated on this card

About the Quadro RTX 8000

The Quadro RTX 8000 runs 27–31B and 35B MoE models well. Turing generation, released 2018.

GenerationTuringreleased 2018

Quadro RTX 8000 strengths and weaknesses

Strengths

  • Runs 35B MoE models: Qwen3.6 35B A3B at 113 tok/s in 22 GB
  • Room for long context: Qwen3.8 27B at Q4 uses 17.9 GB, leaving 30.1 GB free

Weaknesses

  • 120B MoE models don't fit: GPT OSS 120B needs 70.8 GB at Q4 vs 48 GB here
  • Slow prompt reading: A 4K-token prompt on Qwen3.8 27B waits 7.7 s for the first token
  • No FP8 or NVFP4: Turing predates both; use GGUF, AWQ or GPTQ builds

Quadro RTX 8000 tokens per second by model size

Output speed depends on memory bandwidth (672 GB/s). Grey rows need more than 48 GB.

  1. 9BQwen3.5 9B · Q4_K_M · 6.6 GB
    68.7tok/s
  2. 14BQwen3 14B · Q4_K_M · 11.9 GB
    46.0tok/s
  3. 27–31BQwen3.8 27B · Q4_K_M · 17.9 GB
    25.1tok/s
  4. 35B MoEQwen3.6 35B A3B · Q4_K_M · 22.0 GB
    113tok/s
  5. 70BLlama 3.3 70B Instruct · IQ4_XS · 46.4 GB tight · IQ4_XS
    10.5tok/s
  6. 120B MoEGPT OSS 120B · not at 4-bit
    needs 70.8 GB

Estimated speeds: Q4_K_M, 4K context, one user. How we estimate

All 1216 models on the Quadro RTX 8000

Featured only

Best quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works

LLM models compatible with NVIDIA Quadro RTX 8000 — ranked by Score, recency and Fit
ModelVRAMScoreFit
GPT OSS BF16
Q4_K_M·1456.0 t/s
0.3 GBS100, Runs great
Ouro Hybrid 1.4B
BF16·117.1 t/s
6.5 GBS94, Runs great
Jan V3.5 4B
Q4_K_M·139.1 t/s
4.5 GBS96, Runs great
LightOnOCR 3 0.8B
Q4_K_M·507.9 t/s
0.9 GBS100, Runs great
EuroLLM 22B Instruct 2512
Q4_K_M·30.5 t/s
17.5 GBA71, Runs well
Gemma 4 26B A4B
Q4_K_M·126.0 t/s
16.5 GBS95, Runs great
Qwen3.5 9B Claude 4.6 Opus Reasoning Distilled
Q4_K_M·68.7 t/s
6.6 GBS85, Runs great
Qwen3.6 27B PRISM PRO DQ
Q4_K_M·24.5 t/s
21.4 GBA65, Runs well
Tiny Random LlamaForCausalLM
Q4_K_M·1456.0 t/s
0.3 GBS100, Runs great
Falcon H1 7B Instruct
Q4_K_M·86.7 t/s
6.3 GBS89, Runs great
Recursive Language Model 198M
BF16·992.7 t/s
0.4 GBS100, Runs great
GPT OSS 20B Heretic
Q4_K_M·137.9 t/s
13.1 GBS96, Runs great
GLM 4.7 Flash Ultimate Irrefusable Heretic
Q4_K_M·134.5 t/s
19.1 GBS95, Runs great
TeleOCR
Q4_K_M·343.9 t/s
2.1 GBS100, Runs great
Veritas 8B Fact Checker Non Thinking 1.0
Q4_K_M·79.1 t/s
7.6 GBS87, Runs great

Quadro RTX 8000 specs, power and PSU

Specs

Architecture
Turing
VRAM
48 GB GDDR6
Memory bandwidth
672.0 GB/s
FP16 (dense)
65.25 TFLOPS
CUDA cores
4,608
Tensor cores
576
Compute capability
7.5
TDP
260 W
Released
2018-12-01
MSRP
$9,999

Quadro RTX 8000 FAQ

What is the largest model the Quadro RTX 8000 can run?

Qwen2.5 72B Instruct (72.7B) is the largest at 4-bit or better: at Q4_K_S it uses 46.6 GB of 48 GB, and runs at about 10.4 tok/s.

Can the Quadro RTX 8000 run Qwen3.8 27B?

Yes. Qwen3.8 27B fits at Q4_K_M, using 17.9 GB of 48 GB, and it is the best model that runs well on this card.

Can the Quadro RTX 8000 run 120B MoE models?

No. GPT OSS 120B needs 70.8 GB at Q4, more than the 48 GB on this card.

How many tokens per second does the Quadro RTX 8000 give?

About 25.1 tok/s on Qwen3.8 27B at Q4_K_M, with the first token after 7.8 s on a 4K-token prompt.

What power supply does the Quadro RTX 8000 need?

Plan a 550 W PSU (2× the 260 W TDP).

Is 48 GB enough for coding and agents?

Yes for coding: Qwen3.8 27B, the best of the 10 coding models that fit, runs at about 25 tok/s. For agents, not yet: no agent-scored model fits in less than 96 GB.

Related

Best coding LLMs48 GB VRAMGPUs with 48 GB or more