Created bymario-alka.dePowered bygodcore.denoob2claw.detricoma.de
DEEN
Sign in Sign up
Beta
Open data · reproducible tests

Which model is fast & smart?

We benchmark AI language models on real hardware: speed in tokens per second and response time – plus quality on real agent tasks. Transparent, reproducible, directly comparable.

Start typing – or jump straight in:Fastest (tok/s)Best harnessAll modelsMakersHighscore
2853 benchmarks74 models2 test types
Live benchmark
Who generates text the fastest? Real measurements, real hardware.
1
gemma-4-E2B-itGeForce RTX 5090
2.491,2tok/s
2
Nemotron-3-Nano-4BRTX PRO 6000 Blackwe
2.182,7tok/s
3
gpt-oss-20bRTX PRO 6000 Blackwe
1.995,4tok/s
4
Ornith-1.0-35BRTX PRO 6000 Blackwe
1.799,4tok/s
5
Nemotron-3-Nano-Omni-30B-A3B-ReasoningRTX PRO 6000 Blackwe
1.791,5tok/s
Peak2.491tok/s
View ranking →
01

Real performance

TTFT, prefill and generation tokens/s instead of theoretical peak numbers.

02

Harness quality

Points and success rates from complete agent and chat runs.

03

Full transparency

Model, quantization, runtime and hardware stay visible on every run.

Community favorites

Top models by benchmarks

Speed (tokens/s) and quality (harness) of the most-tested models – directly comparable.

All models →
🥇
Qwen (Alibaba)

Qwen3.6-35B-A3B

MoE35B
⚡ Speed1.489,1 tok/s
🏆 Quality–
112 Benchmarks112 Perf.0 Harness
🥈
OpenAI

gpt-oss-20b

MoE20B
⚡ Speed1.995,4 tok/s
🏆 Quality–
110 Benchmarks110 Perf.0 Harness
🥉
MiniMax

MiniMax-M2.5

MoE230B
⚡ Speed551,0 tok/s
🏆 Quality81 %
95 Benchmarks83 Perf.12 Harness
4
Poolside

Laguna-XS-2.1

⚡ Speed1.765,0 tok/s
🏆 Quality–
88 Benchmarks88 Perf.0 Harness
5
Qwen (Alibaba)

Qwen3-30B-A3B-Instruct-2507

MoE30B
⚡ Speed1.671,0 tok/s
🏆 Quality–
79 Benchmarks79 Perf.0 Harness
6
Poolside

Laguna-M.1

⚡ Speed337,6 tok/s
🏆 Quality–
77 Benchmarks77 Perf.0 Harness
Current measurements

Leaderboard Snapshot

Alle Ergebnisse →
Metric:
Size
#Model / MakerMetricsParallelGPU / CPU / RAMRuntime
1gemma-4-E2B-it5BQ4_K_MGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
2.491,20 tok/s TG
Prefill 15.814 · TTFT 4.382 ms
10×NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAMllama.cppcodex_cliDetails →
2gemma-4-E2B-it5BQ4_K_MGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
2.414,37 tok/s TG
Prefill 20.277 · TTFT 4.107 ms
10×NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAMllama.cppgodclawDetails →
3Nemotron-3-Nano-4B4BQ4_K_MNVIDIA Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
2.182,67 tok/s TG
Prefill 10.599 · TTFT 5.598 ms
10×NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAMllama.cppgodclawDetails →
4gemma-4-E2B-it5BQ4_K_MGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
2.143,55 tok/s TG
Prefill 13.632 · TTFT 4.819 ms
10×3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAMllama.cppgodclawDetails →
5gpt-oss-20b20BQ4_K_MOpenAI Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
1.995,37 tok/s TG
Prefill 16.158 · TTFT 5.309 ms
10×NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAMllama.cppgodclawDetails →
Hardware ranking

Benchmarks by graphics card

Generation throughput (tokens/s) per GPU – minimum, average and maximum across all performance runs.

All runs →
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition512 Läufe
555,4tok/s Ø
Min 9,4Max 2.143,6
NVIDIA RTX PRO 6000 Blackwell Workstation Edition244 Läufe
518,2tok/s Ø
Min 0,0Max 2.414,4
NVIDIA GeForce RTX 5090218 Läufe
429,9tok/s Ø
Min 0,1Max 2.491,2
NVIDIA GeForce RTX 3090 Ti264 Läufe
256,4tok/s Ø
Min 0,0Max 1.492,8
NVIDIA GeForce RTX 5070 Ti222 Läufe
167,7tok/s Ø
Min 0,7Max 1.744,0
NVIDIA GB1082 Läufe
148,9tok/s Ø
Min 11,4Max 903,3
Intel Arc Pro B7032 Läufe
141,6tok/s Ø
Min 0,0Max 826,2
AMD Radeon PRO W7900 Dual Slot241 Läufe
106,5tok/s Ø
Min 8,0Max 344,1
NVIDIA GeForce RTX 206039 Läufe
68,6tok/s Ø
Min 6,8Max 133,7
AMD Radeon AI PRO R9700855 Läufe
65,8tok/s Ø
Min 0,0Max 667,0
AMD Radeon 8060S Graphics47 Läufe
61,0tok/s Ø
Min 7,0Max 326,9
NVIDIA Tesla P100 PCIe 16GB28 Läufe
40,0tok/s Ø
Min 13,3Max 111,1
Hardware ranking

Benchmarks by CPU

Generation throughput (tokens/s) per CPU – minimum, average and maximum across all performance runs.

All runs →
AMD Ryzen Threadripper PRO 9965WX 24-Cores510 Läufe
557,4tok/s Ø
Min 9,4Max 2.143,6
AMD Ryzen 9 9950X 16-Core Processor247 Läufe
512,5tok/s Ø
Min 0,0Max 2.414,4
AMD Ryzen 7 5800X3D 8-Core Processor218 Läufe
429,9tok/s Ø
Min 0,1Max 2.491,2
AMD Ryzen 9 8945HX with Radeon Graphics222 Läufe
259,3tok/s Ø
Min 0,0Max 1.492,8
AMD Ryzen 5 5600X 6-Core Processor42 Läufe
241,4tok/s Ø
Min 28,7Max 830,5
NVIDIA Grace81 Läufe
149,8tok/s Ø
Min 11,4Max 903,3
AMD Ryzen Threadripper PRO 5975WX 32-Cores457 Läufe
135,5tok/s Ø
Min 0,7Max 1.744,0
AMD Ryzen 9 7945HX with Radeon Graphics66 Läufe
100,5tok/s Ø
Min 0,0Max 826,2
Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz39 Läufe
68,6tok/s Ø
Min 6,8Max 133,7
AMD Ryzen Threadripper PRO 7955WX 16-Cores855 Läufe
65,8tok/s Ø
Min 0,0Max 667,0
AMD RYZEN AI MAX+ 395 w/ Radeon 8060S47 Läufe
61,0tok/s Ø
Min 7,0Max 326,9
Intel(R) Xeon(R) CPU E5-4620 v2 @ 2.60GHz38 Läufe
4,0tok/s Ø
Min 0,0Max 14,5
Intel(R) Xeon(R) CPU E5-4657L v2 @ 2.40GHz1 Läufe
2,3tok/s Ø
Min 2,3Max 2,3
Frisch gemessen

Neuste Benchmarks

Alle Ergebnisse →
Performance benchmark22.08.2026
Meta-Llama-3.1-8B-Instruct
Meta
45,73 tok/s
Mario Alka
Performance benchmark22.08.2026
Nemotron-3-Nano-4B
NVIDIA
106,92 tok/s
Mario Alka
Performance benchmark22.08.2026
Nemotron-3-Nano-4B
NVIDIA
100,31 tok/s
Mario Alka
Performance benchmark22.08.2026
Nemotron-3-Nano-4B
NVIDIA
72,37 tok/s
Mario Alka
Performance benchmark22.08.2026
Meta-Llama-3.1-8B-Instruct
Meta
48,95 tok/s
Mario Alka
Performance benchmark22.08.2026
Nemotron-3-Nano-4B
NVIDIA
125,68 tok/s
Mario Alka
Performance benchmark22.08.2026
Nemotron-3-Nano-4B
NVIDIA
116,24 tok/s
Mario Alka
Performance benchmark22.08.2026
Nemotron-3-Nano-4B
NVIDIA
80,94 tok/s
Mario Alka
Performance benchmark21.08.2026
Devstral-Small-2-24B-Instruct-2512
Mistral AI
23,29 tok/s
Mario Alka
Performance benchmark21.08.2026
Devstral-Small-2-24B-Instruct-2512
Mistral AI
13,47 tok/s
Mario Alka
Performance benchmark21.08.2026
Devstral-Small-2507
Mistral AI
23,15 tok/s
Mario Alka
Performance benchmark21.08.2026
Devstral-Small-2507
Mistral AI
13,34 tok/s
Mario Alka
Performance benchmark21.08.2026
Magistral-Small-2509
Mistral AI
24,74 tok/s
Mario Alka
Performance benchmark21.08.2026
Magistral-Small-2509
Mistral AI
13,55 tok/s
Mario Alka
Performance benchmark21.08.2026
Mistral-Small-3.1-24B-Instruct-2503
Mistral AI
23,84 tok/s
Mario Alka
Performance benchmark21.08.2026
Mistral-Small-3.1-24B-Instruct-2503
Mistral AI
13,57 tok/s
Mario Alka
Performance benchmark21.08.2026
Codestral-22B-v0.1
Mistral AI
22,24 tok/s
Mario Alka
Performance benchmark21.08.2026
Codestral-22B-v0.1
Mistral AI
13,51 tok/s
Mario Alka
Performance benchmark21.08.2026
gpt-oss-20b
OpenAI
105,30 tok/s
Mario Alka
Performance benchmark21.08.2026
gpt-oss-20b
OpenAI
111,12 tok/s
Mario Alka
Performance benchmark21.08.2026
gpt-oss-20b
OpenAI
63,18 tok/s
Mario Alka
Performance benchmark21.08.2026
Ministral-3-14B-Reasoning-2512
Mistral AI
23,11 tok/s
Mario Alka
Performance benchmark21.08.2026
Ministral-3-14B-Reasoning-2512
Mistral AI
41,46 tok/s
Mario Alka
Performance benchmark21.08.2026
Ministral-3-14B-Reasoning-2512
Mistral AI
22,41 tok/s
Mario Alka
Performance benchmark21.08.2026
gemma-4-12B-it
Google
25,04 tok/s
Mario Alka