AI models, ranked by real usage

What is really running on Elosia: the most used, the best performing, the best value for money.

Most used

118 models tracked
Models34

Top performance

index /100
1Claude Fable 5Claude83
2Claude Opus 4.8Claude82
3Claude Opus 4.7Claude80
4GPT-5.2 CodexGPT80
5Claude Sonnet 5Claude79
6Gemini 3.1 ProGemini78
7Claude Opus 4.6Claude78
8GPT-5.6 Sol ProGPT77
9Kimi K3MoonshotAi76
10GPT-5.2GPT76

Best value

perf ÷ price
1DeepSeek v4 FlashDeepSeek52$0.12
2DeepSeek v3.2DeepSeek61$0.30
3Minimax 2.5Minimax54$0.34
4Z.ai GLM 5.2Z.ai64$0.46
5DeepSeek v3.1DeepSeek54$0.43
6DeepSeek v4 ProDeepSeek60$0.54
7DeepSeek v3 0324DeepSeek53$0.48
8Minimax M3Minimax54$0.52
9Qwen 3.7 PlusQwen55$0.56
10Qwen 3.5 27BQwen55$0.85

By provider

cumulative usage share
Providers8

Favorites by use case

Coding

View all →
  1. 1CClaude Fable 5
  2. 2CClaude Opus 4.8
  3. 3CClaude Opus 4.7

Analysis

View all →
  1. 1CClaude Fable 5
  2. 2CClaude Opus 4.8
  3. 3CClaude Opus 4.7

Research

View all →
  1. 1CClaude Fable 5
  2. 2CClaude Opus 4.8
  3. 3CClaude Opus 4.7

Creative Writing

View all →
  1. 1CClaude Opus 4.8
  2. 2CClaude Opus 4.7
  3. 3CClaude Opus 4.6

Mathematics

View all →
  1. 1GGemini 3.1 Pro
  2. 2GGPT-5.6 Sol Pro
  3. 3GGrok 4.5

Translation

View all →
  1. 1QQwen 3.7 Max
  2. 2QQwen 3.5 27B
  3. 3QQwen 3.6 Plus

Browse all models

14 collections
ProvidersClaudeDeepSeekGeminiGPTGrokMinimaxMistralMoonshotAiQwenZ.ai
TiersFlagshipBalancedCompactSpecialized

Aggregated, anonymized data. Performance is a benchmark index. Prices are real, per 1M tokens. Updated daily.