Created bymario-alka.dePowered bygodcore.denoob2claw.detricoma.de
Contributed byMario AlkaTencent

Tencent-Hy3-295B-A21B

Performance benchmark · measured on 01.08.2026 00:41

Benchmark-IDrun-20260801-072828-43208e
Timebench 3 - Kombi (Prefill + Generation)MoE295BRuntime: llama.cppQuantisierung: Q4_K_M
Generation1,07tok/s
Prefill5,79tok/s
Time to First Token367.448,00ms
Total duration1.200,00s
Concurrency1parallel
Ranking in the field
74of 81 systems

Performance benchmark · Primary metric: Generation-Speed (tok/s) · 1× concurrent

This run is better than 9 % of all comparable systems.
Generation 1,1 tok/s
-99 % vs Ø 152,3
Prefill 5,8 tok/s
-100 % vs Ø 4.182,4
Time to First Token 367.448 ms
+1.333 % vs Ø 25.642
Distribution in the field0 – 395 tok/s
Ø 152 Median Dieser Lauf

Wie schlägt sich dieser Benchmark mit anderen Modellen?

gpt-oss-20bNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260730-035052-9f766a
395,2 tok/s
NVIDIA-Nemotron-3-Nano-4BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260727-181607-bffb11
393,5 tok/s
gemma-4-E2B-itNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260727-181606-6ea089
378,6 tok/s
Nemotron-3-Nano-Omni-30B-A3B-ReasoningNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032119-4852bf
357,1 tok/s
NVIDIA-Nemotron-3-Nano-30B-A3BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032121-643b00
357,0 tok/s
Nemotron-Cascade-2-30B-A3BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032120-4e3476
356,8 tok/s
Qwen3-30B-A3B-Thinking-2507NVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032123-b33938
312,1 tok/s
Qwen3-Coder-30B-A3B-InstructNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032124-ac5167
311,3 tok/s
Laguna-XS-2.1NVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260728-140954-cecd39
310,2 tok/s
Laguna-XS-2.1NVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260728-140955-42efc5
309,9 tok/s
Laguna-XS-2.1NVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260728-140955-abb113
309,5 tok/s
Qwen3-30B-A3B-Instruct-2507NVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032121-b5f531
304,6 tok/s
Qwen3-Omni-30B-A3B-ThinkingNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032125-29dbef
304,4 tok/s
Qwen3-VL-30B-A3B-InstructNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032125-021dfb
303,6 tok/s
Tencent-Hy3-295B-A21B this runNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260801-072828-43208e
1,1 tok/s

How does this benchmark compare on other GPUs?

Same model on different hardware · 1× concurrent · Generation (tok/s)

Hardware

GPU: NVIDIA RTX PRO 6000 Blackwell Workstation Edition · 96 GB VRAM
CPU: AMD Ryzen 9 9950X 16-Core Processor
RAM: 92 GB
Mainboard: ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI

Setup

Runtime: llama.cpp
Quantization: Q4_K_M
Model: Tencent-Hy3-295B-A21B

Configuration

benchmark-konfiguration — run-20260801-072828-43208e
# LLM-Benchmark Konfiguration # Modell : Tencent-Hy3-295B-A21B # Engine : llama.cpp # Run-ID : run-20260801-072828-43208e # GPU : NVIDIA RTX PRO 6000 Blackwell Workstation Edition # CPU : AMD Ryzen 9 9950X 16-Core Processor # RAM : 92 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build/bin/llama-server \ -m /home/godcore/.cache/huggingface/hub/models--AngelSlim--Hy3-GGUF/snapshots/31b453f4d9b647c74e4c4f5cba632eb512332c91/Hy3-Q4_K_M-mtp.gguf \ --alias Tencent-Hy3-295B-A21B \ --host 0.0.0.0 \ --port 8000 \ -ngl 12 \ -c 16384 \ -np 4 \ --jinja
Engine?Die Inferenz-Software, die das Modell ausliefert (z.B. vLLM oder llama.cpp). Sie bestimmt Geschwindigkeit, unterstuetzte Modellformate und welche Parameter ueberhaupt verfuegbar sind.llamacpp
Modellalias?Der Name, unter dem das Modell ueber die API angesprochen wird. Genau dieser Wert muss im Request-Feld 'model' stehen.Tencent-Hy3-295B-A21B
Kontextlaenge?Maximale Anzahl Tokens (Eingabe + erzeugte Ausgabe zusammen), die das Modell pro Anfrage verarbeiten kann.16384
Modellpfad?Pfad zur GGUF-Modelldatei, die geladen und ausgeliefert wird./home/godcore/.cache/huggingface/hub/models--AngelSlim--Hy3-GGUF/snapshots/31b453f4d9b647c74e4c4f5cba632eb512332c91/Hy3-Q4_K_M-mtp.gguf
GPU-Layer?Anzahl der auf die GPU ausgelagerten Modell-Layer. Hoeher = mehr VRAM und schneller; der Rest laeuft auf der CPU. 999 = alles auf GPU.12
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.16384
np4

All benchmarks of this model To leaderboard

Anzeige
Model comparison

Tencent-Hy3-295B-A21B on various hardware

All published performance runs of this model – each bubble a variant: position = prefill (X) × generation (Y), bubble size = number of runs. Closer to the top right = faster. ★ Marked gold = this benchmark.

GPUby graphics card

4073062041020,005101.0211.531Prefill (tok/s)Generation (tok/s)NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition - 313,7 tok/s Generation, 1.236 tok/s Prefill, TTFT 45.437 ms (9 Laufe)NVIDIA RTX PRO 6000 B...NVIDIA GeForce RTX 5070 Ti - 5,5 tok/s Generation, 49 tok/s Prefill, TTFT 134.636 ms (3 Laufe)NVIDIA GeForce RTX 50...NVIDIA RTX PRO 6000 Blackwell Workstation Edition - 1,7 tok/s Generation, 8 tok/s Prefill, TTFT 324.394 ms (4 Laufe) | DIESER LAUF★ NVIDIA RTX PRO 6000 B...
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 313,7 tok/sNVIDIA GeForce RTX 5070 Ti 5,5 tok/s★ NVIDIA RTX PRO 6000 Blackwell Workstation Edition 1,7 tok/s this run

CPUby processor

4073062041020,005101.0211.531Prefill (tok/s)Generation (tok/s)AMD Ryzen Threadripper PRO 9965WX 24-Cores - 313,7 tok/s Generation, 1.236 tok/s Prefill, TTFT 45.437 ms (9 Laufe)AMD Ryzen Threadrippe...AMD Ryzen Threadripper PRO 5975WX 32-Cores - 5,5 tok/s Generation, 49 tok/s Prefill, TTFT 134.636 ms (3 Laufe)AMD Ryzen Threadrippe...AMD Ryzen 9 9950X 16-Core Processor - 1,7 tok/s Generation, 8 tok/s Prefill, TTFT 324.394 ms (4 Laufe) | DIESER LAUF★ AMD Ryzen 9 9950X 16-...
AMD Ryzen Threadripper PRO 9965WX 24-Cores 313,7 tok/sAMD Ryzen Threadripper PRO 5975WX 32-Cores 5,5 tok/s★ AMD Ryzen 9 9950X 16-Core Processor 1,7 tok/s this run

MBby mainboard

4073062041020,005101.0211.531Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 313,7 tok/s Generation, 1.236 tok/s Prefill, TTFT 45.437 ms (9 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 5,5 tok/s Generation, 49 tok/s Prefill, TTFT 134.636 ms (3 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI - 1,7 tok/s Generation, 8 tok/s Prefill, TTFT 324.394 ms (4 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 313,7 tok/sASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 5,5 tok/s★ ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 1,7 tok/s this run

ENGby engine

345329314298282664692721749Prefill (tok/s)Generation (tok/s)llama.cpp - 313,7 tok/s Generation, 706 tok/s Prefill, TTFT 131.901 ms (16 Laufe) | DIESER LAUF★ llama.cpp
★ llama.cpp 313,7 tok/s this run

DRVby driver

345329314298282664692721749Prefill (tok/s)Generation (tok/s)unbekannt - 313,7 tok/s Generation, 706 tok/s Prefill, TTFT 131.901 ms (16 Laufe)unbekannt
unbekannt 313,7 tok/s
💰 Economics

Economics of this run

Operating cost, TCO and comparison with the next-best runs of the same model at identical concurrency (1× concurrent). Methodology →

⚙️ ConfigurationAll metrics and charts below follow these settings – based on a 24-month runtime.Save to URLReset
⚡ Electricity price EUR/kWh
⚙️ System utilization 100 %
🖥️ Acquisition EUR
🔌 Idle 70 W
⚡ TDP 644 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)644 W estimated (TDP)GPU 600 + CPU 29 + Board 15 W full load
Avg cost / hourEUR 0.19
Electricity / 1M tokensEUR 50.14
Token / kWh5.98K
Acquisition (system)EUR 15,248 full priceGPU EUR 13,000 · CPU EUR 649 · Board EUR 499 · RAM EUR 920 · PSU EUR 180
Electricity (2 years)
TCO (2 years)EUR 18,632
Output tokens (2 years)67.49M
☁️ External LLM (API) – comparison
External LLM cost (2 years)
Savings vs. external (2 years)

All values above and the charts below take the configured system utilization into account: at X% the system generates only X% of the time, the rest it idles (70 W). Cost per hour drops (more idle), cost per token rises.

Cost over 2 years – electricity only

Cost over 2 years – incl. acquisition (TCO)

Speed vs. tokens per euro

Euro per 1M tokens

Comparison vs. API – economics per benchmark

Tencent-Hy3-295B-A21BNVIDIA RTX PRO 6000 Blackwell Workstation EditionTencent-Hy3-295B-A21B3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionTencent-Hy3-295B-A21B3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionTencent-Hy3-295B-A21B3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
Electricity cost (24 mo.)
Acquisition cost
Total cost (TCO)
Generated tokens (24 mo.)
Token price via API
Break-even point (days)
Result (savings / extra cost)

Comparison with up to 3 next-best runs of this model at the same concurrency (at least one on different hardware). Power = GPU TDP + CPU (idle + 15 %) + board (estimated), acquisition = full system (GPU + CPU + board + RAM + PSU), prices = stored market prices.

Contributed by

Mario Alka Administrator

@marioalka

Ich bin Unternehmer, Softwareentwickler und KI-Enthusiast. Seit vielen Jahren entwickle ich Unternehmenssoftware und beschäftige mich inzwischen fast täglich mit lokalen LLMs, KI-Agenten und leistungsfähiger KI-Hardware.

Mit LLM-Benchmark.de möchte ich eine Plattform schaffen, auf der Modelle, GPUs und Agenten objektiv und reproduzierbar miteinander verglichen werden.