LLMs im direkten Vergleich 7 Einsatzszenarien · Gleiche Bedingungen · eine Ergebnisliste

Wie gut schlagen sich freie Open-Weight-Modelle gegen proprietäre API-Modelle und lizenzlich eingeschränkte Restricted-Weight-Modelle? Anhand alltäglicher Aufgaben und unter denselben Bedingungen landen alle Lizenzmodelle und Gewichtsklassen im selben Scoreboard.

Der Gesamtscore zeigt den besten Allrounder. Die Profile darunter beantworten die eigentliche Alltagsfrage: Reicht dieses Modell für deine Aufgabe – auch wenn es ein kleiner, lokaler Spezialist ist statt ein Cloud-Gigant?

#							Report
1	GLM-5.3 Restricted Server Cloud	82.3	Thinking	80Tool-Use unterstützt, Score: 80	137 t/s	$1.57	Report
Module Scores		Code 84.8 CLI 95.3 Reason. 77.2 UX-W. 79.8 Doc. 80.5 Content 86.0 Cultural 81.3 Synth. 73.3 Tool Ex. 90.0
2	Claude Opus 5.5 Commercial Frontier API	81.0	Thinking	77Tool-Use unterstützt, Score: 77	0.5 t/s	$1.36	Report
Module Scores		Code 88.9 CLI 86.7 Reason. 75.8 UX-W. 81.6 Doc. 85.7 Content 83.0 Cultural 72.8 Synth. 62.5 Tool Ex. 90.0
3	Qwen 3.8 Flash-Next (Thinking) Open Weight Server vLLM	81.0	Thinking	85Tool-Use unterstützt, Score: 85	41.2 t/s	–	Report
Module Scores		Code 81.3 CLI 86.0 Reason. 77.7 UX-W. 79.8 Doc. 82.2 Content 74.3 Cultural 84.9 Synth. 62.5 Tool Ex. 90.0
4	GLM-5.3-Flash (EXL3, TensorFold) Open Weight Server TensorFold	80.9	Thinking	74Tool-Use unterstützt, Score: 74	55.5 t/s	–	Report
Module Scores		Code 85.1 CLI 94.3 Reason. 80.0 UX-W. 79.1 Doc. 82.5 Content 82.9 Cultural 76.0 Synth. 73.3 Tool Ex. 90.0
5	Claude Opus 4.8 Commercial Frontier API	80.6	Adaptive	78Tool-Use unterstützt, Score: 78	0.3 t/s	$1.14	Report
Module Scores		Code 84.2 CLI 97.7 Reason. 77.7 UX-W. 77.4 Doc. 79.7 Content 83.5 Cultural 75.6 Synth. 66.7 Tool Ex. 90.0
6	Swift Qwen 3.8 27B (Thinking) Restricted Workstation vLLM	80.5	Thinking	75Tool-Use unterstützt, Score: 75	16.3 t/s	–	Report
Module Scores		Code 84.4 CLI 90.7 Reason. 75.3 UX-W. 81.4 Doc. 83.3 Content 82.2 Cultural 76.8 Synth. 59.2 Tool Ex. 90.0
7	GLM-5.3-Flash Open Weight Server Cloud	79.7	Thinking	80Tool-Use unterstützt, Score: 80	79.4 t/s	$0.14	Report
Module Scores		Code 81.3 CLI 87.3 Reason. 76.9 UX-W. 79.4 Doc. 79.7 Content 78.9 Cultural 77.7 Synth. 68.3 Tool Ex. 90.0
8	Qwen 3.8 27B Uncensored (Thinking) Open Weight Workstation vLLM	79.7	Thinking	79Tool-Use unterstützt, Score: 79	16.6 t/s	–	Report
Module Scores		Code 79.7 CLI 93.7 Reason. 78.3 UX-W. 75.3 Doc. 80.1 Content 79.8 Cultural 78.2 Synth. 59.2 Tool Ex. 90.0
9	Kimi K3 Open Weight Frontier Cloud	79.3	Thinking	81Tool-Use unterstützt, Score: 81	91.4 t/s	$3.23	Report
Module Scores		Code 84.4 CLI 83.7 Reason. 78.2 UX-W. 73.8 Doc. 78.3 Content 75.7 Cultural 82.1 Synth. 73.3 Tool Ex. 88.3
10	Claude Opus 5 Commercial Frontier API	78.7	Thinking	83Tool-Use unterstützt, Score: 83	0.3 t/s	$1.56	Report
Module Scores		Code 81.9 CLI 80.9 Reason. 77.2 UX-W. 74.6 Doc. 86.0 Content 81.3 Cultural 66.4 Synth. 80.0 Tool Ex. 83.3
11	Xiaomi MiMo V2.5 Pro Open Weight Frontier Cloud	78.7	Adaptive	73Tool-Use unterstützt, Score: 73	32.7 t/s	$0.11	Report
Module Scores		Code 79.8 CLI 89.0 Reason. 74.0 UX-W. 75.9 Doc. 86.4 Content 83.2 Cultural 73.8 Synth. 44.2 Tool Ex. 90.0
12	Xiaomi MiMo V2.6 Pro Open Weight Frontier Cloud	78.3	Standard	75Tool-Use unterstützt, Score: 75	27.1 t/s	$0.14	Report
Module Scores		Code 78.4 CLI 82.0 Reason. 81.7 UX-W. 75.0 Doc. 83.2 Content 78.1 Cultural 75.0 Synth. 59.2 Tool Ex. 90.0
13	Qwen3.8-Flash Open Weight Server Cloud	78.2	Thinking	80Tool-Use unterstützt, Score: 80	60.8 t/s	$0.05	Report
Module Scores		Code 85.3 CLI 88.3 Reason. 77.7 UX-W. 75.2 Doc. 79.1 Content 70.4 Cultural 74.9 Synth. 62.5 Tool Ex. 90.0
14	Qwen 3.8 27B (Thinking) Open Weight Workstation vLLM	78.2	Thinking	79Tool-Use unterstützt, Score: 79	21.3 t/s	–	Report
Module Scores		Code 87.7 CLI 89.0 Reason. 77.2 UX-W. 76.0 Doc. 74.8 Content 72.8 Cultural 74.6 Synth. 59.2 Tool Ex. 90.0
15	Meta Muse Spark 1.2 Commercial Frontier Cloud	77.7	Adaptive	78Tool-Use unterstützt, Score: 78	162 t/s	$0.54	Report
Module Scores		Code 75.3 CLI 89.0 Reason. 77.3 UX-W. 73.1 Doc. 72.0 Content 80.5 Cultural 82.4 Synth. 0.0 Tool Ex. 0.0
16	Claude Sonnet 4.6 Commercial Frontier API	77.7	Adaptive	67Tool-Use unterstützt, Score: 67	0.2 t/s	$0.66	Report
Module Scores		Code 78.1 CLI 94.3 Reason. 79.9 UX-W. 80.0 Doc. 78.9 Content 76.5 Cultural 74.9 Synth. 51.7 Tool Ex. 83.3
17	Kimi K2.6 Open Weight Frontier Cloud	77.4	Adaptive	82Tool-Use unterstützt, Score: 82	51.8 t/s	$1.05	Report
Module Scores		Code 77.6 CLI 90.7 Reason. 72.5 UX-W. 71.1 Doc. 75.4 Content 79.1 Cultural 77.4 Synth. 63.3 Tool Ex. 86.7
18	Claude Haiku 5.5 Commercial Frontier API	77.3	Thinking	77Tool-Use unterstützt, Score: 77	197 t/s	$0.09	Report
Module Scores		Code 84.9 CLI 88.3 Reason. 76.2 UX-W. 72.7 Doc. 70.7 Content 79.6 Cultural 74.5 Synth. 66.7 Tool Ex. 90.0
19	MiniMax M3 Open Weight Server Cloud	77.3	Adaptive	80Tool-Use unterstützt, Score: 80	85.3 t/s	$0.16	Report
Module Scores		Code 78.8 CLI 85.8 Reason. 75.5 UX-W. 75.8 Doc. 79.5 Content 76.0 Cultural 71.0 Synth. 62.5 Tool Ex. 90.0
20	Gemini 3.5 Flash Commercial Frontier API	77.2	Adaptive	76Tool-Use unterstützt, Score: 76	58.7 t/s	$0.49	Report
Module Scores		Code 74.8 CLI 93.0 Reason. 73.9 UX-W. 76.9 Doc. 74.3 Content 73.3 Cultural 82.8 Synth. 63.3 Tool Ex. 83.3
21	Claude Sonnet 5 Commercial Frontier API	76.7	Thinking	79Tool-Use unterstützt, Score: 79	0.4 t/s	$0.46	Report
Module Scores		Code 78.4 CLI 90.7 Reason. 74.5 UX-W. 76.0 Doc. 75.8 Content 74.6 Cultural 72.1 Synth. 66.7 Tool Ex. 83.3
22	GPT-5.5 Commercial Frontier API	76.7	Thinking	78Tool-Use unterstützt, Score: 78	81.6 t/s	$2.90	Report
Module Scores		Code 79.0 CLI 86.7 Reason. 71.3 UX-W. 75.3 Doc. 72.3 Content 81.7 Cultural 74.5 Synth. 62.5 Tool Ex. 83.3
23	Kimi K2.5 Open Weight Frontier Cloud	76.6	Thinking	77Tool-Use unterstützt, Score: 77	37.7 t/s	$0.39	Report
Module Scores		Code 77.8 CLI 88.3 Reason. 75.3 UX-W. 71.7 Doc. 79.4 Content 70.9 Cultural 78.4 Synth. 65.8 Tool Ex. 90.0
24	Qwen3.8-2.4T-A95B Open Weight Frontier Cloud	76.5	Thinking	81Tool-Use unterstützt, Score: 81	95.0 t/s	$2.22	Report
Module Scores		Code 82.0 CLI 86.7 Reason. 69.8 UX-W. 72.8 Doc. 72.5 Content 77.3 Cultural 74.6 Synth. 66.7 Tool Ex. 90.0
25	GLM 4.6 Restricted Server Cloud	76.2	Standard	53Tool-Use unterstützt, Score: 53	46.1 t/s	$0.25	Report
Module Scores		Code 72.6 CLI 89.0 Reason. 71.6 UX-W. 71.3 Doc. 74.4 Content 80.2 Cultural 81.3 Synth. 63.3 Tool Ex. 89.2
26	Ornith 1.5 35B-A3B Open Weight Workstation vLLM	76.0	Thinking	76Tool-Use unterstützt, Score: 76	61.0 t/s	–	Report
Module Scores		Code 79.3 CLI 91.3 Reason. 73.7 UX-W. 70.6 Doc. 78.7 Content 80.0 Cultural 65.8 Synth. 60.0 Tool Ex. 71.7
27	Qwen 3.7 Max Commercial Frontier Cloud	75.8	Adaptive	60Tool-Use unterstützt, Score: 60	62.6 t/s	$0.75	Report
Module Scores		Code 77.8 CLI 93.0 Reason. 76.2 UX-W. 72.4 Doc. 74.9 Content 81.7 Cultural 78.9 Synth. 63.3 Tool Ex. 76.7
28	Ornith 1.0 35B Open Weight Workstation vLLM	75.8	Thinking	63Tool-Use unterstützt, Score: 63	35.5 t/s	–	Report
Module Scores		Code 76.9 CLI 90.7 Reason. 74.0 UX-W. 77.2 Doc. 78.7 Content 71.8 Cultural 81.4 Synth. 62.5 Tool Ex. 80.8
29	Gemini 3.8 Flash Commercial Frontier Cloud	75.7	Thinking	83Tool-Use unterstützt, Score: 83	114 t/s	$0.40	Report
Module Scores		Code 71.6 CLI 86.0 Reason. 71.8 UX-W. 76.4 Doc. 74.3 Content 75.2 Cultural 73.2 Synth. 76.7 Tool Ex. 90.0
30	Qwen 3.6 27B (Thinking) Open Weight Workstation vLLM	75.6	Thinking	73Tool-Use unterstützt, Score: 73	25.2 t/s	–	Report
Module Scores		Code 70.9 CLI 93.0 Reason. 75.8 UX-W. 73.4 Doc. 76.4 Content 78.8 Cultural 71.8 Synth. 49.2 Tool Ex. 71.7
31	GLM-5.2 Open Weight Server Cloud	75.5	Adaptive	51Tool-Use unterstützt, Score: 51	65.7 t/s	$0.40	Report
Module Scores		Code 70.3 CLI 84.3 Reason. 76.7 UX-W. 74.0 Doc. 78.6 Content 78.4 Cultural 73.6 Synth. 55.8 Tool Ex. 90.0
32	Qwen 3.6 Plus Commercial Frontier Cloud	75.5	Adaptive	68Tool-Use unterstützt, Score: 68	56.3 t/s	$0.41	Report
Module Scores		Code 73.8 CLI 89.0 Reason. 73.6 UX-W. 73.3 Doc. 74.8 Content 80.0 Cultural 77.8 Synth. 50.0 Tool Ex. 68.3
33	Upstage Solar Pro4 Commercial Frontier Cloud	75.4	Standard	71Tool-Use unterstützt, Score: 71	35.9 t/s	$0.01	Report
Module Scores		Code 74.4 CLI 89.9 Reason. 74.8 UX-W. 76.1 Doc. 73.2 Content 78.7 Cultural 72.9 Synth. 66.7 Tool Ex. 90.0
34	Qwen 3.8 Omni Flash Commercial Frontier Cloud	75.3	Thinking	78Tool-Use unterstützt, Score: 78	42.1 t/s	$0.19	Report
Module Scores		Code 81.8 CLI 85.0 Reason. 69.0 UX-W. 73.3 Doc. 71.7 Content 70.8 Cultural 77.8 Synth. 66.7 Tool Ex. 90.0
35	Occamy 1.0 35B-A3B (Accio-Lab) (Thinking) Open Weight Workstation vLLM	75.3	Thinking	70Tool-Use unterstützt, Score: 70	33.9 t/s	–	Report
Module Scores		Code 74.1 CLI 83.3 Reason. 74.5 UX-W. 77.0 Doc. 74.2 Content 78.2 Cultural 74.6 Synth. 42.5 Tool Ex. 86.7
36	Xiaomi MiMo V2.6 Flash Open Weight Server Cloud	75.0	Standard	73Tool-Use unterstützt, Score: 73	43.2 t/s	$0.03	Report
Module Scores		Code 78.2 CLI 90.7 Reason. 75.6 UX-W. 69.9 Doc. 79.8 Content 70.0 Cultural 70.6 Synth. 59.2 Tool Ex. 90.0
37	Gemma 4 26B-A4B Instruct (Thinking) Open Weight Workstation vLLM	74.9	Thinking	74Tool-Use unterstützt, Score: 74	27.3 t/s	–	Report
Module Scores		Code 70.1 CLI 89.0 Reason. 76.5 UX-W. 68.5 Doc. 75.2 Content 78.6 Cultural 74.2 Synth. 60.0 Tool Ex. 82.5
38	Claude Haiku 4.5 Commercial Frontier API	74.8	Standard	67Tool-Use unterstützt, Score: 67	98.4 t/s	$0.43	Report
Module Scores		Code 81.2 CLI 93.7 Reason. 75.6 UX-W. 75.0 Doc. 69.2 Content 73.3 Cultural 73.2 Synth. 52.5 Tool Ex. 83.3
39	Mistral 3 Large Open Weight Server API	74.8	Standard	69Tool-Use unterstützt, Score: 69	50.7 t/s	$0.10	Report
Module Scores		Code 77.6 CLI 82.7 Reason. 74.0 UX-W. 77.0 Doc. 75.6 Content 75.8 Cultural 71.0 Synth. 30.0 Tool Ex. 90.0
40	DeepSeek V4.1 Flash Open Weight Server Cloud	74.8	Thinking	80Tool-Use unterstützt, Score: 80	99.9 t/s	$0.12	Report
Module Scores		Code 74.6 CLI 93.0 Reason. 65.5 UX-W. 70.1 Doc. 68.5 Content 77.9 Cultural 78.1 Synth. 63.3 Tool Ex. 90.0
41	Grok 4.6 Commercial Frontier API	74.7	Thinking	73Tool-Use unterstützt, Score: 73	13.6 t/s	$0.44	Report
Module Scores		Code 81.1 CLI 88.3 Reason. 64.4 UX-W. 73.1 Doc. 74.1 Content 69.7 Cultural 80.6 Synth. 54.0 Tool Ex. 90.0
42	DeepSeek-V4.1-Flash (EXL3) (Thinking) Open Weight Server vLLM	74.5	Thinking	78Tool-Use unterstützt, Score: 78	30.5 t/s	–	Report
Module Scores		Code 78.2 CLI 78.0 Reason. 68.1 UX-W. 66.0 Doc. 72.6 Content 79.6 Cultural 77.4 Synth. 66.7 Tool Ex. 90.0
43	Ministral 3 14B (Unsloth) Open Weight Desktop llama.cpp	74.5	Standard	61Tool-Use unterstützt, Score: 61	15.8 t/s	–	Report
Module Scores		Code 77.1 CLI 81.1 Reason. 70.3 UX-W. 74.0 Doc. 76.6 Content 77.8 Cultural 81.1 Synth. 33.3 Tool Ex. 85.8
44	Gemini 3.7 Flash Commercial Frontier Cloud	74.4	Adaptive	80Tool-Use unterstützt, Score: 80	115 t/s	$0.34	Report
Module Scores		Code 72.2 CLI 93.0 Reason. 67.2 UX-W. 72.5 Doc. 73.5 Content 76.3 Cultural 70.0 Synth. 70.0 Tool Ex. 90.0
45	Occamy 1.0 35B-A3B (Accio-Lab) Open Weight Workstation vLLM	74.4	Standard	69Tool-Use unterstützt, Score: 69	30.0 t/s	–	Report
Module Scores		Code 71.5 CLI 88.3 Reason. 74.3 UX-W. 71.5 Doc. 73.5 Content 78.4 Cultural 75.7 Synth. 59.2 Tool Ex. 87.5
46	GPT-OSS 120B (Thinking) Open Weight Server vLLM	74.4	Thinking	76Tool-Use unterstützt, Score: 76	31.4 t/s	–	Report
Module Scores		Code 75.0 CLI 90.7 Reason. 73.6 UX-W. 74.6 Doc. 70.5 Content 71.5 Cultural 71.0 Synth. 55.0 Tool Ex. 90.0
47	Qwen 3.5 35B-A3B (Unsloth) Open Weight Workstation llama.cpp	74.3	Adaptive	72Tool-Use unterstützt, Score: 72	69.3 t/s	–	Report
Module Scores		Code 72.5 CLI 91.7 Reason. 70.4 UX-W. 74.3 Doc. 70.6 Content 76.4 Cultural 75.6 Synth. 55.8 Tool Ex. 90.0
48	Kimi K2.7 Code Open Weight Frontier Cloud	74.3	Thinking	63Tool-Use unterstützt, Score: 63	83.0 t/s	$0.57	Report
Module Scores		Code 74.9 CLI 82.0 Reason. 75.6 UX-W. 70.1 Doc. 75.5 Content 78.5 Cultural 78.4 Synth. 70.0 Tool Ex. 90.0
49	Mistral Medium 3.5 Open Weight Server API	74.2	Standard	71Tool-Use unterstützt, Score: 71	106 t/s	$0.46	Report
Module Scores		Code 74.8 CLI 90.7 Reason. 69.6 UX-W. 76.6 Doc. 67.1 Content 76.9 Cultural 75.3 Synth. 59.2 Tool Ex. 90.0
50	Qwen 3 Coder Next Open Weight Server llama.cpp	74.1	Standard	71Tool-Use unterstützt, Score: 71	48.5 t/s	–	Report
Module Scores		Code 79.8 CLI 88.9 Reason. 64.6 UX-W. 71.7 Doc. 68.1 Content 75.7 Cultural 80.3 Synth. 51.7 Tool Ex. 90.0
51	Qwen 3.6 35B-A3B (Unsloth) (Thinking) Open Weight Workstation vLLM	74.1	Thinking	77Tool-Use unterstützt, Score: 77	84.6 t/s	–	Report
Module Scores		Code 70.6 CLI 67.7 Reason. 74.2 UX-W. 73.1 Doc. 74.8 Content 82.0 Cultural 70.4 Synth. 56.7 Tool Ex. 83.3
52	GLM-4.7 Restricted Server Cloud	74.0	Adaptive	71Tool-Use unterstützt, Score: 71	47.7 t/s	$0.25	Report
Module Scores		Code 67.1 CLI 86.7 Reason. 73.5 UX-W. 76.3 Doc. 74.7 Content 74.9 Cultural 73.8 Synth. 42.5 Tool Ex. 83.3
53	Grok 4.5 Commercial Frontier API	73.8	Standard	77Tool-Use unterstützt, Score: 77	25.6 t/s	$0.52	Report
Module Scores		Code 77.2 CLI 90.7 Reason. 67.7 UX-W. 64.4 Doc. 79.2 Content 72.4 Cultural 70.6 Synth. 60.0 Tool Ex. 90.0
54	GPT 5.6 Luna Commercial Frontier API	73.6	Thinking	80Tool-Use unterstützt, Score: 80	87.6 t/s	$0.10	Report
Module Scores		Code 80.5 CLI 89.0 Reason. 59.8 UX-W. 67.2 Doc. 67.8 Content 77.2 Cultural 75.3 Synth. 75.0 Tool Ex. 87.5
55	Gemma 4 26B-A4B Q5_K_M (ARA-Abliterated) Open Weight Workstation llama.cpp	73.3	Standard	76Tool-Use unterstützt, Score: 76	53.5 t/s	–	Report
Module Scores		Code 74.9 CLI 90.0 Reason. 68.2 UX-W. 67.3 Doc. 68.4 Content 72.3 Cultural 77.6 Synth. 63.3 Tool Ex. 90.0
56	GLM-5.1 Open Weight Server Cloud	73.3	Adaptive	68Tool-Use unterstützt, Score: 68	63.5 t/s	$0.52	Report
Module Scores		Code 71.3 CLI 93.0 Reason. 74.4 UX-W. 72.2 Doc. 69.9 Content 79.5 Cultural 68.5 Synth. 59.2 Tool Ex. 90.0
57	Mistral Small 4 Open Weight Server API	73.3	Standard	20Tool-Use unterstützt, Score: 20	117 t/s	$0.03	Report
Module Scores		Code 70.9 CLI 79.7 Reason. 72.5 UX-W. 72.3 Doc. 76.4 Content 72.6 Cultural 71.4 Synth. 39.2 Tool Ex. 66.7
58	Signal 3.8 27B Open Weight Workstation llama.cpp	73.2	Standard	77Tool-Use unterstützt, Score: 77	14.5 t/s	–	Report
Module Scores		Code 64.6 CLI 93.7 Reason. 69.2 UX-W. 74.4 Doc. 65.2 Content 75.2 Cultural 76.6 Synth. 66.7 Tool Ex. 90.0
59	Gemma 4 31B Instruct Open Weight Workstation Cloud	73.2	Standard	74Tool-Use unterstützt, Score: 74	30.6 t/s	$0.02	Report
Module Scores		Code 73.6 CLI 89.0 Reason. 73.3 UX-W. 68.7 Doc. 64.5 Content 79.1 Cultural 72.4 Synth. 60.0 Tool Ex. 90.0
60	Qwen 3.5 397B A17B Open Weight Server Cloud	73.2	Adaptive	74Tool-Use unterstützt, Score: 74	20.4 t/s	$0.46	Report
Module Scores		Code 72.8 CLI 90.7 Reason. 72.7 UX-W. 63.8 Doc. 67.2 Content 78.5 Cultural 74.2 Synth. 60.0 Tool Ex. 90.0
61	Gemini 2.5 Pro Commercial Frontier API	73.2	Adaptive	71Tool-Use unterstützt, Score: 71	29.7 t/s	$0.68	Report
Module Scores		Code 74.4 CLI 86.7 Reason. 70.1 UX-W. 71.5 Doc. 68.2 Content 76.4 Cultural 73.7 Synth. 60.0 Tool Ex. 90.0
62	MiniMax M2.7 Restricted Server Cloud	73.0	Adaptive	65Tool-Use unterstützt, Score: 65	48.0 t/s	$0.13	Report
Module Scores		Code 74.5 CLI 89.0 Reason. 70.3 UX-W. 76.2 Doc. 72.4 Content 75.5 Cultural 69.2 Synth. 43.3 Tool Ex. 57.5
63	Gemma 4 12B Instruct (Unsloth, Q6_K_XL) — Instruct-Profil Open Weight Desktop llama.cpp	72.9	Standard	75Tool-Use unterstützt, Score: 75	13.3 t/s	–	Report
Module Scores		Code 79.2 CLI 92.2 Reason. 65.4 UX-W. 68.5 Doc. 62.4 Content 72.0 Cultural 78.0 Synth. 60.0 Tool Ex. 90.0
64	DeepSeek V4 Flash Open Weight Server Cloud	72.9	Adaptive	78Tool-Use unterstützt, Score: 78	29.9 t/s	$0.01	Report
Module Scores		Code 71.4 CLI 86.7 Reason. 70.8 UX-W. 65.6 Doc. 68.7 Content 78.6 Cultural 71.7 Synth. 66.7 Tool Ex. 90.0
65	Gemma 4 12B Instruct (Unsloth, Q6_K_XL) Open Weight Desktop llama.cpp	72.7	Thinking	71Tool-Use unterstützt, Score: 71	16.6 t/s	–	Report
Module Scores		Code 74.8 CLI 85.6 Reason. 67.7 UX-W. 71.8 Doc. 63.9 Content 75.2 Cultural 77.6 Synth. 56.7 Tool Ex. 88.3
66	DeepSeek V4 Pro Open Weight Frontier Cloud	72.6	Adaptive	77Tool-Use unterstützt, Score: 77	53.3 t/s	$0.30	Report
Module Scores		Code 74.8 CLI 83.7 Reason. 65.7 UX-W. 71.3 Doc. 63.8 Content 76.1 Cultural 73.8 Synth. 63.3 Tool Ex. 90.0
67	NVIDIA Nemotron 3 Ultra 550B A55B Open Weight Server Cloud	72.5	Adaptive	55Tool-Use unterstützt, Score: 55	97.0 t/s	$0.37	Report
Module Scores		Code 79.1 CLI 95.3 Reason. 72.5 UX-W. 75.7 Doc. 73.1 Content 70.1 Cultural 71.0 Synth. 29.2 Tool Ex. 83.3
68	Gemini 3.5 Flash Lite Commercial Frontier Cloud	72.5	Thinking	74Tool-Use unterstützt, Score: 74	164 t/s	$0.14	Report
Module Scores		Code 66.1 CLI 84.3 Reason. 77.0 UX-W. 69.1 Doc. 66.2 Content 74.9 Cultural 74.5 Synth. 60.0 Tool Ex. 88.3
69	Gemma 4 E4B Open Weight Edge llama.cpp	71.9	Adaptive	73Tool-Use unterstützt, Score: 73	48.2 t/s	–	Report
Module Scores		Code 71.0 CLI 81.1 Reason. 70.9 UX-W. 70.7 Doc. 64.7 Content 72.6 Cultural 75.6 Synth. 56.7 Tool Ex. 90.0
70	Qwen 3.6 35B-A3B (Uncensored) Open Weight Workstation llama.cpp	71.7	Adaptive	60Tool-Use unterstützt, Score: 60	53.9 t/s	–	Report
Module Scores		Code 74.3 CLI 90.6 Reason. 67.1 UX-W. 72.0 Doc. 73.3 Content 70.5 Cultural 75.6 Synth. 40.8 Tool Ex. 81.7
71	Muse Glimmer 30B Open Weight Workstation vLLM	71.6	Thinking	73Tool-Use unterstützt, Score: 73	10.6 t/s	–	Report
Module Scores		Code 77.0 CLI 90.0 Reason. 60.1 UX-W. 66.8 Doc. 66.9 Content 68.1 Cultural 80.3 Synth. 60.0 Tool Ex. 85.8
72	Ministral 3 8B (Unsloth) Open Weight Edge llama.cpp	71.3	Standard	60Tool-Use unterstützt, Score: 60	24.9 t/s	–	Report
Module Scores		Code 73.9 CLI 79.5 Reason. 68.8 UX-W. 76.3 Doc. 69.0 Content 77.1 Cultural 70.0 Synth. 29.2 Tool Ex. 89.2
73	Grok 4.7 Commercial Frontier API	71.1	Thinking	72Tool-Use unterstützt, Score: 72	17.4 t/s	$0.56	Report
Module Scores		Code 69.5 CLI 90.7 Reason. 59.0 UX-W. 73.2 Doc. 62.1 Content 74.2 Cultural 77.8 Synth. 53.3 Tool Ex. 90.0
74	Qwen 3.5 27B Open Weight Workstation vLLM	71.0	Thinking	66Tool-Use unterstützt, Score: 66	17.8 t/s	–	Report
Module Scores		Code 64.3 CLI 83.7 Reason. 65.1 UX-W. 69.2 Doc. 71.0 Content 80.8 Cultural 74.6 Synth. 52.5 Tool Ex. 83.3
75	Xiaomi MiMo V2.5 Open Weight Server Cloud	70.4	Adaptive	71Tool-Use unterstützt, Score: 71	43.0 t/s	$0.03	Report
Module Scores		Code 68.6 CLI 86.7 Reason. 72.8 UX-W. 68.3 Doc. 62.9 Content 71.4 Cultural 70.2 Synth. 70.0 Tool Ex. 90.0
76	Devstral 2 Open Weight Server API	70.2	Standard	55Tool-Use unterstützt, Score: 55	56.2 t/s	$0.13	Report
Module Scores		Code 67.8 CLI 80.3 Reason. 72.9 UX-W. 72.5 Doc. 70.2 Content 80.3 Cultural 67.2 Synth. 43.3 Tool Ex. 69.2
77	GPT-5.4 Mini Commercial Frontier API	70.2	Standard	68Tool-Use unterstützt, Score: 68	120 t/s	$0.25	Report
Module Scores		Code 74.4 CLI 81.0 Reason. 65.7 UX-W. 63.5 Doc. 61.5 Content 77.7 Cultural 75.3 Synth. 55.8 Tool Ex. 83.3
78	Qwen 3.5 4B (Unsloth) Open Weight Nano llama.cpp	69.9	Adaptive	78Tool-Use unterstützt, Score: 78	51.3 t/s	–	Report
Module Scores		Code 63.5 CLI 72.8 Reason. 67.5 UX-W. 71.8 Doc. 68.6 Content 73.2 Cultural 64.6 Synth. 70.0 Tool Ex. 87.5
79	NVIDIA Nemotron 3.5 Lightning 30B (Thinking) Open Weight Workstation vLLM	69.6	Thinking	73Tool-Use unterstützt, Score: 73	93.5 t/s	–	Report
Module Scores		Code 64.5 CLI 95.0 Reason. 69.1 UX-W. 68.7 Doc. 58.7 Content 65.8 Cultural 74.6 Synth. 55.8 Tool Ex. 90.0
80	o4-mini Commercial Frontier API	69.5	Thinking	63Tool-Use unterstützt, Score: 63	69.7 t/s	$0.40	Report
Module Scores		Code 75.3 CLI 90.6 Reason. 56.0 UX-W. 64.3 Doc. 66.2 Content 73.0 Cultural 78.3 Synth. 40.8 Tool Ex. 85.0
81	GPT-5.4 Commercial Frontier API	69.4	Standard	43Tool-Use unterstützt, Score: 43	76.9 t/s	$0.89	Report
Module Scores		Code 77.7 CLI 88.3 Reason. 67.1 UX-W. 73.0 Doc. 68.3 Content 75.3 Cultural 72.0 Synth. 55.0 Tool Ex. 61.7
82	Laguna S 2.1 Open Weight Server vLLM	69.1	Thinking	63Tool-Use unterstützt, Score: 63	17.2 t/s	$0.02	Report
Module Scores		Code 67.2 CLI 91.3 Reason. 67.9 UX-W. 67.4 Doc. 62.9 Content 77.0 Cultural 67.6 Synth. 33.3 Tool Ex. 78.3
83	Hermes 4 70B Open Weight Server Cloud	69.0	Adaptive	76Tool-Use unterstützt, Score: 76	76.8 t/s	$0.03	Report
Module Scores		Code 65.2 CLI 78.0 Reason. 73.0 UX-W. 62.1 Doc. 65.1 Content 63.1 Cultural 74.4 Synth. 41.7 Tool Ex. 88.3
84	Gemma 3 12B IT Restricted Desktop llama.cpp	68.9	Standard	64Tool-Use unterstützt, Score: 64	39.1 t/s	–	Report
Module Scores		Code 69.5 CLI 87.2 Reason. 59.9 UX-W. 64.5 Doc. 62.0 Content 75.8 Cultural 77.3 Synth. 48.3 Tool Ex. 83.3
85	Qwen 3.5 9B (Unsloth) Open Weight Edge llama.cpp	68.7	Adaptive	57Tool-Use unterstützt, Score: 57	34.3 t/s	–	Report
Module Scores		Code 70.8 CLI 77.8 Reason. 70.7 UX-W. 65.5 Doc. 65.7 Content 80.7 Cultural 65.6 Synth. 46.7 Tool Ex. 68.3
86	Hermes 4 405B Restricted Server Cloud	68.4	Adaptive	69Tool-Use unterstützt, Score: 69	31.1 t/s	$0.17	Report
Module Scores		Code 68.0 CLI 80.3 Reason. 61.6 UX-W. 61.6 Doc. 71.0 Content 73.5 Cultural 68.6 Synth. 60.0 Tool Ex. 90.0
87	NVIDIA Nemotron 3 Nano 30B A3B Open Weight Workstation Cloud	68.3	Adaptive	66Tool-Use unterstützt, Score: 66	35.4 t/s	$0.02	Report
Module Scores		Code 65.9 CLI 86.7 Reason. 67.3 UX-W. 61.2 Doc. 66.2 Content 73.9 Cultural 64.6 Synth. 51.7 Tool Ex. 82.5
88	Hermes 4 14B (Abliterated) Open Weight Desktop llama.cpp	68.2	Adaptive	70Tool-Use unterstützt, Score: 70	25.4 t/s	–	Report
Module Scores		Code 62.9 CLI 85.6 Reason. 64.5 UX-W. 61.5 Doc. 64.6 Content 74.4 Cultural 71.0 Synth. 36.7 Tool Ex. 90.0
89	Gemma 4 E2B (Unsloth) Open Weight Edge llama.cpp	68.1	Thinking	59Tool-Use unterstützt, Score: 59	67.1 t/s	–	Report
Module Scores		Code 71.4 CLI 85.6 Reason. 69.4 UX-W. 63.5 Doc. 58.3 Content 74.2 Cultural 72.3 Synth. 50.0 Tool Ex. 68.3
90	GPT-5.4 Nano Commercial Frontier API	68.1	Standard	49Tool-Use unterstützt, Score: 49	125 t/s	$0.07	Report
Module Scores		Code 71.4 CLI 78.7 Reason. 67.0 UX-W. 69.5 Doc. 67.8 Content 79.7 Cultural 67.0 Synth. 46.7 Tool Ex. 67.5
91	o3-mini Commercial Frontier API	68.0	Thinking	73Tool-Use unterstützt, Score: 73	71.2 t/s	$0.37	Report
Module Scores		Code 64.9 CLI 87.8 Reason. 52.9 UX-W. 64.2 Doc. 60.4 Content 73.0 Cultural 78.3 Synth. 55.8 Tool Ex. 90.0
92	Hermes 4.3 36B (Thinking) Open Weight Server vLLM	68.0	Thinking	67Tool-Use unterstützt, Score: 67	13.3 t/s	–	Report
Module Scores		Code 67.8 CLI 86.1 Reason. 63.3 UX-W. 62.0 Doc. 66.9 Content 72.2 Cultural 67.3 Synth. 44.2 Tool Ex. 90.0
93	Qwen 3 14B Open Weight Desktop llama.cpp	67.7	Adaptive	66Tool-Use unterstützt, Score: 66	23.9 t/s	–	Report
Module Scores		Code 67.8 CLI 91.7 Reason. 63.1 UX-W. 62.5 Doc. 60.9 Content 70.4 Cultural 71.3 Synth. 41.7 Tool Ex. 90.0
94	Hermes 4 14B Open Weight Desktop llama.cpp	67.3	Adaptive	67Tool-Use unterstützt, Score: 67	30.3 t/s	–	Report
Module Scores		Code 62.5 CLI 85.6 Reason. 64.0 UX-W. 56.6 Doc. 64.8 Content 73.3 Cultural 73.6 Synth. 45.0 Tool Ex. 90.0
95	Hermes 4.3 36B Open Weight Server vLLM	67.1	Standard	67Tool-Use unterstützt, Score: 67	12.7 t/s	–	Report
Module Scores		Code 66.8 CLI 82.2 Reason. 62.2 UX-W. 60.9 Doc. 62.6 Content 69.4 Cultural 74.0 Synth. 43.3 Tool Ex. 90.0
96	Ornith 1.0 9B (Unsloth) Open Weight Desktop llama.cpp	65.9	Thinking	71Tool-Use unterstützt, Score: 71	25.1 t/s	–	Report
Module Scores		Code 66.1 CLI 80.3 Reason. 66.5 UX-W. 56.9 Doc. 56.8 Content 61.2 Cultural 75.6 Synth. 52.5 Tool Ex. 90.0
97	Codestral 25.08 Restricted Desktop API	65.5	Standard	69Tool-Use unterstützt, Score: 69	192 t/s	$0.05	Report
Module Scores		Code 66.3 CLI 84.3 Reason. 64.1 UX-W. 60.8 Doc. 62.7 Content 64.2 Cultural 62.0 Synth. 56.7 Tool Ex. 83.3
98	Ministral 3 3B (Unsloth) Open Weight Nano llama.cpp	64.8	Standard	56Tool-Use unterstützt, Score: 56	53.0 t/s	–	Report
Module Scores		Code 65.3 CLI 75.0 Reason. 62.1 UX-W. 68.0 Doc. 68.2 Content 71.8 Cultural 56.3 Synth. 32.5 Tool Ex. 83.3
99	GPT-OSS 20B Open Weight Desktop vLLM	64.3	Standard	28Tool-Use unterstützt, Score: 28	41.5 t/s	–	Report
Module Scores		Code 78.9 CLI 87.2 Reason. 66.2 UX-W. 69.2 Doc. 61.2 Content 66.5 Cultural 69.0 Synth. 20.0 Tool Ex. 35.0
100	Llama 8B (Unsloth, provenance unverified) Restricted Edge llama.cpp	62.7	Standard	Kein Tool-Use	25.5 t/s	–	Report
Module Scores		Code 54.5 CLI 82.2 Reason. 56.5 UX-W. 59.3 Doc. 54.7 Content 63.3 Cultural 78.3 Synth. n/a Tool Ex. n/a
101	Gemma 3 4B (Unsloth) Restricted Nano llama.cpp	61.5	Standard	Kein Tool-Use	44.5 t/s	–	Report
Module Scores		Code 58.8 CLI 71.1 Reason. 58.7 UX-W. 60.1 Doc. 47.1 Content 68.5 Cultural 71.0 Synth. n/a Tool Ex. n/a
102	Qwen 3 4B Open Weight Nano llama.cpp	60.9	Adaptive	71Tool-Use unterstützt, Score: 71	74.0 t/s	–	Report
Module Scores		Code 57.1 CLI 73.3 Reason. 58.0 UX-W. 58.6 Doc. 62.8 Content 64.8 Cultural 48.3 Synth. 51.7 Tool Ex. 90.0
103	Hermes 3 8B Restricted Edge llama.cpp	58.8	Standard	56Tool-Use unterstützt, Score: 56	48.0 t/s	–	Report
Module Scores		Code 56.7 CLI 80.6 Reason. 48.8 UX-W. 59.1 Doc. 48.1 Content 64.2 Cultural 67.6 Synth. 30.0 Tool Ex. 83.3
104	Phi-4 Mini (Unsloth) Open Weight Nano llama.cpp	58.4	Standard	66Tool-Use unterstützt, Score: 66	46.1 t/s	–	Report
Module Scores		Code 56.7 CLI 78.9 Reason. 61.1 UX-W. 57.8 Doc. 50.9 Content 53.7 Cultural 52.3 Synth. 47.5 Tool Ex. 88.3
105	DeepSeek R1 Distill Qwen 14B Open Weight Desktop llama.cpp	58.3	Standard	Kein Tool-Use	15.0 t/s	–	Report
Module Scores		Code 62.6 CLI 71.7 Reason. 48.3 UX-W. 57.8 Doc. 57.6 Content 62.7 Cultural 53.9 Synth. n/a Tool Ex. n/a
106	Qwen 2.5 Coder 7B Open Weight Edge llama.cpp	55.3	Standard	58Tool-Use unterstützt, Score: 58	37.4 t/s	–	Report
Module Scores		Code 52.3 CLI 85.6 Reason. 53.4 UX-W. 54.8 Doc. 50.2 Content 58.9 Cultural 44.6 Synth. 34.2 Tool Ex. 82.5
107	Llama 3.2 3B (Unsloth) Restricted Nano llama.cpp	52.0	Standard	48Tool-Use unterstützt, Score: 48	55.0 t/s	–	Report
Module Scores		Code 46.8 CLI 70.0 Reason. 45.0 UX-W. 54.1 Doc. 43.3 Content 61.0 Cultural 57.3 Synth. 31.7 Tool Ex. 63.3
108	Llama 3.2 1B (Unsloth) Restricted Nano llama.cpp	44.4	Standard	Kein Tool-Use	130 t/s	–	Report
Module Scores		Code 28.1 CLI 71.1 Reason. 38.3 UX-W. 40.0 Doc. 36.6 Content 60.1 Cultural 49.6 Synth. n/a Tool Ex. n/a
109	DeepSeek R1 Distill Qwen 7B Open Weight Edge llama.cpp	42.6	Standard	Kein Tool-Use	29.3 t/s	–	Report
Module Scores		Code 32.0 CLI 57.8 Reason. 41.4 UX-W. 45.3 Doc. 40.5 Content 43.5 Cultural 45.3 Synth. n/a Tool Ex. n/a
110	DeepSeek R1 Distill Qwen 1.5B Open Weight Nano llama.cpp	35.4	Standard	Kein Tool-Use	110 t/s	–	Report
Module Scores		Code 31.4 CLI 45.0 Reason. 33.4 UX-W. 31.1 Doc. 35.4 Content 40.1 Cultural 36.2 Synth. n/a Tool Ex. n/a
111	Gemma 3 270M (Unsloth) Restricted Nano llama.cpp	23.6	Standard	Kein Tool-Use	211 t/s	–	Report
Module Scores		Code 15.2 CLI 51.1 Reason. 17.4 UX-W. 19.9 Doc. 10.0 Content 33.8 Cultural 31.4 Synth. n/a Tool Ex. n/a