LIVE TELEMETRYGPT-5$1.25 / $10.00Sonnet 5.5$2.00 / $10.00Gemini 3.8$0.75 / $3.75DS V4.1$0.15 / $0.60RTX 50901,792 GB/sM4 Max546 GB/sUS Avg$0.182/kWh
TELEMETRY CONSOLE NAVIGATION

Guncelkal Telemetry Index

Search VRAM specs, API costs, or Dev tools

AI & LLM EconomicsQ5_K_M • 19.25 GB Total

Mistral Small 24B Q6_K Fit Check on RTX 4090

High-precision 6-bit reasoning without crossing the 24 GB GDDR6X limit.

Hardware & Model Configuration

16,384 tokens
1 stream
64 GB
KV Cache Precision:
CLI LAUNCH SNIPPET
OLLAMA_FLASH_ATTENTION=1 OLLAMA_KV_CACHE_TYPE=q8_0 \
  ollama run mistral-small:24b-instruct-q5_k_m --ctx-size 16384
Total Required
19.25 GB
Safe cap: 22.8 GB
Est. Speed
42 tok/s
Decode bandwidth bound
KV Cache
1.25 GB
Q8_0 @ 16K ctx
Max Context Fit
32,768
Tokens before spill
VRAM ALLOCATION & KV-CACHE MAP19.3 GB / 22.8 GB Usable
Safe Pool (22.8 GB)Physical 24 GB
Weights: 17.1 GB KV Cache: 1.25 GB Runtime: 0.9 GB
Status: Fits Wholly in GPU VRAM (Full Speed)

All 7 GGUF / Quantization Tiers on NVIDIA RTX 4090 (24 GB)

QuantizationMemory NeededSpeed Est.Placement
FP16 / BF16 (16-bit)Reference Lossless50.15 GB1.5 tok/sRAM Offload
Q8_0 (8-bit)Near-Lossless27.65 GB4 tok/sRAM Offload
Q6_K (6-bit K-Quant)Studio Fidelity21.83 GB32.9 tok/sPure GPU VRAM
Q5_K_M (5-bit Medium)Minimal Loss19.25 GB42 tok/sPure GPU VRAM
Q4_K_M (4-bit Sweet Spot)Recommended Default16.7 GB49.2 tok/sPure GPU VRAM
IQ3_XS (3.3-bit I-Quant)Noticeable Degradation12.35 GB69.9 tok/sPure GPU VRAM
IQ2_XXS (2.1-bit Extreme)Emergency Fit Only8.75 GB106.9 tok/sPure GPU VRAM

All 11 Open-Weight LLMs Ranked for NVIDIA RTX 4090 (24 GB) (Click Row to Switch)

ModelBest Pure-VRAM QuantMemoryEst. Speed
Llama 3.1 8B InstructUltra-fast RAG extraction, IDE autocomplete & classificationFP1617.96 GB44.8 tok/s
Qwen 2.5 Coder 14BMid-tier local VS Code copilot balancing speed & repo accuracyQ8_018.02 GB45.8 tok/s
Mistral Small 24BClean instruction following on 24 GB single-GPU workstationsQ6_K21.83 GB32.9 tok/s
Gemma 2 / 3 27BMultilingual & editorial writing in single 24–32 GB GPUsQ4_K_M20.26 GB42.9 tok/s
Qwen 2.5 32B / DeepSeek-R1 32BGold-standard local coding & reasoning on RTX 3090/4090/5090 & 48GB MacsQ4_K_M22.6 GB32.7 tok/s
Mixtral 8x7B MoE (46.7B Total)Sparse Mixture-of-Experts: needs 47B RAM footprint, computes at 13B speedIQ3_XS22.05 GB116 tok/s
Llama 3.3 70B / DeepSeek-R1 70BGPT-4 class private local reasoning for 48 GB–128 GB rigsNeeds Offload / Cluster46.5 GB1.8 tok/s
DeepSeek R1 671B MoE (Full Frontier)Cluster or Mac Studio Ultra / dual EPYC server deployment (37B active per token)Needs Offload / Cluster409.9 GBOOM
Qwen3-30B-A3B (MoE — 3.3B Active)Ultra-fast 30B Sparse MoE decoding at 3.3B speed; effortless fit on 24 GB cardsQ4_K_M20.14 GB343.5 tok/s
Qwen3-32B (Dense Dual-Thinking)2026 dense local reasoning & coding benchmark with dual thinking/non-thinking toggleQ4_K_M22.78 GB32.4 tok/s
Llama 4 Scout 109B (MoE — 17B Active)Multimodal 109B MoE (17B active) built for 64–128 GB Macs, Dual-3090, or H100 podsNeeds Offload / Cluster68.78 GB6 tok/s

Telemetry Methodology & Benchmark Sources

Read full equations →

Pre-configured telemetry for Mistral Small 24B (Q5_K_M) on NVIDIA RTX 4090 (24 GB). Requires 19.25 GB (17.1 GB weights + 1.25 GB Q8 KV cache) against 22.8 GB usable pool for ~42 tok/s decode.

Primary Technical & US Grid Benchmarks

  • llama.cpp GGUF Model Sizing
  • Apple Metal / CUDA Wired VRAM Limits
  • GQA Attention Context Scaling Spec

Engineering & Pricing FAQ

Can NVIDIA RTX 4090 (24 GB) run Mistral Small 24B without CPU RAM offload?
24B at Q5_K_M delivers ~70+ tok/s on RTX 4090. At Q5_K_M with 16K context, total required memory is 19.25 GB compared to 22.8 GB safe usable GPU VRAM.
What max context length fits inside NVIDIA RTX 4090 (24 GB)?
At Q5_K_M quantization on single-stream inference, safe KV-cache capacity supports up to 32,768 tokens before spilling into host system RAM.

Continue in Console Workflow

AI & LLM Economics

Cloud vs. Rig TCO

Plot cumulative 36-month API spending against hourly RunPod/H100 rentals and owned RTX/Mac hardware with electricity and resale value.

Open Workbench
Hardware & Power

kWh Power Cost Lab

Calculate daily, monthly, and annual kWh bills using 80+ PSU efficiency, duty cycle, optional AC cooling load, and 50 US state EIA rates.

Open Workbench
Hardware & Power

PSU Wattage & Spike

Combine dual-GPU draw, CPU peak boost, NVMe arrays, and ATX 3.1 transient spikes to hit the 50–70% PSU efficiency sweet spot.

Open Workbench