🏆 Top 25 OSS — Artificial Analysis

Fusion Tiny (≤4B) + Small (4B-40B) + Medium (40B-150B) + Large (>150B) — classement par Efficacité (Intelligence / Total Params)

Scrapé le dimanche 30 août 2026 à 18:36:51 UTC

25
Modèles OSS
Gemma 4 26B A4B
Meilleur OSS
26.07
Score Top
1.2182
Efficacité Top
#8
Créateurs

Classement unifié Top 25 Open Source par Efficacité

Intelligence / Total Params — plus haut = + efficace (MoE + Dense)
#ModèleCréateurIntelligenceParamètresContexteEfficacité
1 Gemma 4 26B A4B Google 26.07 Active: 3.8B / Total: 21.4B 256K 1.2182
2 G9v3-39A5B – 34.02 Active: 5.0B / Total: 34.0B 131K 1.0006
3 Qwen3.6 35B A3B Alibaba 32.13 Active: 3.0B / Total: 33.0B 262K 0.9736
4 Qwen3.6 35B A3B (Non-reasoning) Alibaba 24.55 Active: 3.0B / Total: 33.0B 262K 0.7439
5 HyperNova 60B 2605 – 18.32 Active: 4.8B / Total: 53.9B 131K 0.3399
6 Qwen3.8-Flash-Next Alibaba 55.81 Active: 6.0B / Total: 174.0B 256K 0.3207
7 Ling 3.0 Flash ByteDance 37.82 Active: 5.1B / Total: 118.9B 262K 0.3181
8 Qwen3.5 122B A10B Alibaba 32.85 Active: 10.0B / Total: 115.0B 262K 0.2857
9 Qwen3 Coder Next Alibaba 21.29 Active: 3.0B / Total: 76.7B 256K 0.2776
10 LongCat Flash Lite LongCat 17.39 Active: 3.0B / Total: 65.5B 256K 0.2655
11 Qwen3.5 122B A10B (Non-reasoning) Alibaba 28.18 Active: 10.0B / Total: 115.0B 262K 0.245
12 Nemotron 3 Super NVIDIA 25.67 Active: 12.7B / Total: 107.9B 1.0M 0.2379
13 gpt-oss-120b (high) Google 24.13 Active: 5.1B / Total: 111.9B 131K 0.2156
14 DeepSeek V4 Flash 0731 (max) – 51.77 Active: 13.0B / Total: 271.0B 1.0M 0.191
15 GLM-5.3-Flash – 57.46 Active: 18.0B / Total: 302.0B 1.0M 0.1903
16 Mistral Small 4 Mistral AI 19.71 Active: 6.5B / Total: 112.5B 256K 0.1752
17 Motif 3 – 47.36 Active: 13.2B / Total: 300.8B 262K 0.1574
18 MiniMax-M3 – 45.4 Active: 23.0B / Total: 405.0B 1.0M 0.1121
19 GLM-5.3 (max) – 59.51 Active: 40.0B / Total: 713.0B 1.0M 0.0835
20 GLM-5.2 (max) – 52.64 Active: 40.0B / Total: 713.0B 1.0M 0.0738
21 DeepSeek V4 Pro 0813 (max) – 53.2 Active: 49.0B / Total: 1551.0B 1.0M 0.0343
22 DeepSeek V4 Pro (max) – 45.27 Active: 49.0B / Total: 1551.0B 1.0M 0.0292
23 Qwen3.8 2.4T A95B Alibaba 57.7 Active: 95.0B / Total: 2305.0B 984K 0.025
24 Kimi K3 (max) MBZUAI 59.7 Active: 104.0B / Total: 2696.0B 1.0M 0.0221
25 Kimi K3 (low) MBZUAI 48.25 Active: 104.0B / Total: 2676.0B 1.0M 0.018