Builds / RTX 5090 32GB

RTX 5090: Flint 40B NVFP4 at 224.9 tok/s, 110 tok/s at full 262K context (llama.cpp b11803)

Name: RTX 5090: Flint 40B NVFP4 at 224.9 tok/s, 110 tok/s at full 262K context (llama.cpp b11803)
Creator: RigBase
License: https://creativecommons.org/licenses/by/4.0/

Source: niklaslenz.com ↗📏 How measured: reported by the original source; RigBase re-checked the numbers at that source

GPU(s): RTX 5090 32GB
Amazon · eBay (used) · Newegg
VRAM / unified memory (GB): 32
CPU: not stated
Motherboard: not stated
RAM: not stated
PSU: not stated
Other: Official llama.cpp b11803 prebuilt (ubuntu-cuda-12.8-x64, includes Flint PR #30928), maker's llama-server flags (-c 262144, -np 1, -fa on, f16 KV). Flint always reasons (no thinking switch). Peak GPU memory 28.76 GiB at 262K context
Reported total cost: not stated
Power: not stated

Measured speed (generation tokens/sec)

Model	Quant	Engine	tok/s	Prompt processing (tok/s)	Details
Flint 40B-A2.5B	NVFP4 GGUF (RecursiveSI/flint-nvfp4-gguf), f16 KV	llama.cpp b11803, no speculation	224.9	6,263 @ 8,192-token prompt est. TTFT @8k ≈ 1.3 s (calculated from measured prefill) measured TTFT: 1.3 s @ 8,192-token prompt	context: 262,144 ctx, 512 generated tokens📏 Mean of the code (224.82) and prose (224.94) medians; 2 runs each at 512 tokens after 1 warm-up, measured 2026-10-22
Flint 40B-A2.5B	NVFP4 GGUF, f16 KV	llama.cpp b11803, no speculation	110.2	2,355 @ 249,036-token prompt est. TTFT @8k ≈ 3.5 s (calculated from measured prefill)	context: 262,144 ctx, 249,036-token prompt📏 Memory test: decode on the answer at the largest context (prompt fills 95% of 262,144). Peak GPU memory 28.76 GiB

Prompt processing = prefill speed (how fast the prompt is read) as reported by the source, with the prompt length it was measured at; blank = the source does not report it. est. TTFT @8k is calculated from the measured prefill speed (8,192 ÷ prefill tok/s), shown only when prefill was measured on a prompt of 512+ tokens. It is not a measurement and ignores load time and caching. measured TTFT is the source's own time-to-first-token. Why prefill matters for coding assistants.

niklaslenz.com home benchmark (Results v2), measured 2026-10-22. CPU, board, RAM and PSU not stated.

Current GPU street price (estimate, not the builder's cost)

1× RTX 5090: ≈€7,140 each (≈$8,003), new, approx. price as of 2026-10-24, lowest listing on billiger.de (MSI GeForce RTX 5090 (lowest variant))

Estimated GPU cost for this build today: ≈€7,140
New retail prices in Germany incl. VAT; used cards are often much cheaper. Prices change daily.

Source: https://niklaslenz.com/v2/flint-40b-5090-nvfp4-llamacpp/

Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.

Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)

Quick answers: Can RTX 5090 run Flint 40B-A2.5B?