News-Tracker

Modeller

AI & LLM · daglig briefing

Ændringer siden forrige snapshot

Score ændret

  • Gemini 3.1 Pro Preview: intelligence 38.3 → 36.7 (-1.6).
  • Mistral Large 3: intelligence 9.7 → 11.1 (+1.4).

Ny i HF-trending

  • Qwen/Qwen-Drive-1.0-4B er ny i Hugging Faces trending (score 44).
  • Viggle/Viggle-Animate er ny i Hugging Faces trending (score 44).
  • inclusionAI/Ling-3.0-tiny er ny i Hugging Faces trending (score 39).

Intelligence Index — Artificial Analysis

#ModelIntel.Agentic CreatorUdgivet
1 Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) 5 effort-varianter 56.8 58.2 Anthropic 2026-09-01
2 GPT-6 Astra (max) 6 effort-varianter 54.7 51.6 OpenAI 2026-09-03
3 Claude Opus 5 (Adaptive Reasoning, Max Effort) 5 effort-varianter 54.1 56.4 Anthropic 2026-07-24
4 Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) 53.2 51.2 Anthropic 2026-06-09
5 Muse Spark 1.3 (max) 2 effort-varianter 53 55.6 Meta 2026-09-02
6 GPT-5.6 Sol (max) 6 effort-varianter 51.3 50.7 OpenAI 2026-07-09
7 Grok 4.6 (high) 4 effort-varianter 50.6 53.6 SpaceXAI 2026-08-12
8 Kimi K3 (max) 2 effort-varianter ⬇ hf.co/moonshotai/Kimi-K3 50.2 50.9 Kimi 2026-07-16
9 GLM-5.3 (max) ⬇ hf.co/zai-org/GLM-5.3 48.6 53.6 Z AI 2026-08-18
10 Gemini 3.8 Flash (high) 3 effort-varianter 47.1 41.2 Google 2026-09-02
11 Qwen3.8 Max 46.9 49.9 Alibaba 2026-08-03
12 GPT-5.6 Terra (max) 6 effort-varianter 46.8 43.9 OpenAI 2026-07-09
13 Muse Spark 1.2 (xhigh) 46.8 44.2 Meta 2026-08-05
14 Qwen3.8 2.4T A95B ⬇ hf.co/Qwen/Qwen3.8-2.4T-A95B 46.7 50.7 Alibaba 2026-08-12
15 Claude Opus 4.8 (Adaptive Reasoning, Max Effort) 46.4 42.8 Anthropic 2026-05-28
16 GLM-5.3-Flash ⬇ hf.co/zai-org/GLM-5.3-Flash 46.2 51.5 Z AI 2026-08-26
17 GPT-5.5 (xhigh) 5 effort-varianter 45.6 37.5 OpenAI 2026-04-23
18 Qwen3.8-Flash-Next ⬇ hf.co/Qwen/Qwen3.8-Flash-Next 45.6 Alibaba 2026-08-26
19 Grok 4.5 (high) 45.5 42.4 SpaceXAI 2026-07-08
20 Gemini 3.7 Flash (high) 3 effort-varianter 45.2 36.6 Google 2026-08-13
21 Claude Sonnet 5 (Adaptive Reasoning, Max Effort) 6 effort-varianter 45.1 44.5 Anthropic 2026-06-30
22 Claude Opus 4.7 (Adaptive Reasoning, Max Effort) 2 effort-varianter 44.3 39.7 Anthropic 2026-04-16
23 GPT-5.6 Luna (max) 6 effort-varianter 43.4 42.9 OpenAI 2026-07-09
24 Muse Spark 1.1 (xhigh) 43.3 27.8 Meta 2026-07-09
25 GPT-5.4 (xhigh) 3 effort-varianter 42.8 OpenAI 2026-03-05
26 GLM-5.2 (max) 2 effort-varianter ⬇ hf.co/zai-org/GLM-5.2 42.5 39.7 Z AI 2026-06-16
27 DeepSeek V4 Pro 0813 (Reasoning, Max Effort) ⬇ hf.co/deepseek-ai/DeepSeek-V4-Pro 42.1 42.5 DeepSeek 2026-08-13
28 Gemini 3.5 Flash (high) 3 effort-varianter 41.9 27.4 Google 2026-05-19
29 DeepSeek V4 Flash Vision (Reasoning, Max Effort) 41.5 DeepSeek 2026-08-21
30 Qwen3.8 27B (xhigh) 4 effort-varianter ⬇ hf.co/Qwen/Qwen3.8-27B 41.4 46.8 Alibaba 2026-08-14
31 DeepSeek V4 Flash 0731 (Reasoning, Max Effort) ⬇ hf.co/deepseek-ai/DeepSeek-V4-Flash-0731 40.8 41.9 DeepSeek 2026-07-31
32 Gemini 3.6 Flash (high) 40.3 30.3 Google 2026-07-21
33 Agnes 2.5 Pro Beta 39.5 Sapiens AI 2026-08-26
34 Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) 3 effort-varianter 38.5 33.3 Anthropic 2026-02-17
35 Motif 3 2 effort-varianter ⬇ hf.co/Motif-Technologies/Motif-3 38 Motif Technologies 2026-08-12
36 K2 Horizon 375B A23B ⬇ hf.co/IFM/K2-Horizon-375B-A23B 37.8 41.2 MBZUAI Institute of Foundation Models 2026-09-03
37 Qwen3.7 Max 37 23.9 Alibaba 2026-05-19
38 GPT-5.3 Codex (xhigh) 36.9 OpenAI 2026-02-05
39 Gemini 3.1 Pro Preview 36.7 10.4 Google 2026-02-19
40 Claude Opus 4.6 (Adaptive Reasoning, Max Effort) 2 effort-varianter 36.4 Anthropic 2026-02-05

Hugging Face — trending familier

#Rod-repoTrendingLikes DownloadsParamsAlder
1 deepseek-ai/DeepSeek-V4-Flash-Vision-Exp +1 derivater 613 682 184542 304.6B 6 d
2 Qwen/Qwen3.8-27B +17 derivater 535 14042 6024467 27.8B 32 d
3 Qwen/Qwen3.8-Flash-Next +5 derivater 472 4918 401327 180.0B 13 d
4 XHToken/Spark-X2.5-4B-Base +2 derivater 455 549 4755 4.1B 13 d
5 google/timesfm-3.0-pytorch 401 458 123025 330.7M 12 d
6 zai-org/GLM-5.3-Flash +3 derivater 376 2081 727610 321.3B 12 d
7 Lightricks/LTX-2.5 365 2893 1484329 45 d
8 zai-org/GLM-5.3 +1 derivater 342 1720 370417 753.3B 12 d
9 openai-community/gpt2 254 3702 14739982 137.0M 1648 d
10 nreimers/MiniLM-L6-H384-uncased +1 derivater 254 5559 255006933 22.7M 1648 d
11 BreezeBlue/Breeze-TTS-2 245 449 5962 3.5B 11 d
12 MiniMaxAI/MiniMax-H3 +10 derivater 232 4937 5057414 33.1B 40 d
13 google-bert/bert-base-uncased 221 2986 56175564 110.1M 1648 d
14 facebook/mms-300m 217 263 12961 1202 d
15 distilbert/distilbert-base-uncased 205 1156 7101423 67.0M 1648 d

Params er safetensors.total, altså lagrede vægte inklusive kvantiserede tensorer — ikke et mål for hvad modellen kræver af hardware.