8 guides · 130 models · recomputed every 10 minutes

Which model for which need

Short rankings, recomputed from comparator data, to start from a real constraint — hardware, language, budget, hosting — rather than an abstract benchmark.

• 24 Go56 models

  Best local LLM for 24 GB VRAM

  1. 1Gemma 4 31B1451
  2. 2Gemma 4 26B-A4B1438
  3. 3Gemma 3 27B1365

  See the ranking
• 8 Go30 models

  Best local LLM for a laptop (8 GB)

  1. 1Gemma 3n E4B1317
  2. 2Granite 4.1 8B1306
  3. 3Gemma 3 4B1303

  See the ranking
• FR40 models

  Best LLMs for French

  1. 1Qwen3.7 Max1474
  2. 2GLM-51458
  3. 3Qwen3.5 397B-A17B1441

  See the ranking
• UE20 models

  Best European AI models

  1. 1Mistral Large 31414
  2. 2Codestral1240
  3. 3Ministral 8B1237

  See the ranking
• code34 models

  Best self-hosted coding LLM

  1. 1Kimi K2.61461
  2. 2GLM-51458
  3. 3Gemma 4 31B1451

  See the ranking
• 128k+78 models

  Best LLM for enterprise RAG

  1. 1Claude Opus 4.71502
  2. 2Gemini 3.7 Flash1491
  3. 3Gemini 3 Pro1486

  See the ranking
• ≤ 2 $12 models

  Best API LLMs under $2 per million output tokens

  1. 1GLM-51458
  2. 2Qwen3.7 Plus1455
  3. 3DeepSeek V3.21425

  See the ranking
• agents48 models

  Best LLMs for AI agents

  1. 1Claude Fable 51507
  2. 2Claude Opus 51492
  3. 3Muse Spark 1.11492

  See the ranking