Model Leaderboard
BetaCompare 20+ AI models by quality, speed, and cost on real Xora usage data.
23
Models Ranked
1.2M
Total API Calls
4,820
USDC Volume
Live
Data
| 1 | C Claude 3.5 SonnetAnthropic | 98.2 | 71 | $0.004 | A+ |
| 2 | G GPT-4oOpenAI | 96.1 | 65 | $0.005 | A |
| 3 | G Gemini 2.5 FlashGoogle | 94.5 | 88 | $0.001 | A |
| 4 | C Claude 3 OpusAnthropic | 93.4 | 28 | $0.015 | A |
| 5 | G Gemini 2.5 ProGoogle | 92.6 | 54 | $0.007 | A |
| 6 | L Llama 3.3 70BMeta · Open | 91.0 | 57 | $0.0009 | B+ |
| 7 | G GPT-4 TurboOpenAI | 89.7 | 42 | $0.010 | B+ |
| 8 | D DeepSeek V3DeepSeek · Open | 88.5 | 63 | $0.0006 | B+ |
| 9 | M Mixtral 8x22BMistral · Open | 87.3 | 49 | $0.0008 | B |
| 10 | G Grok 4xAI | 86.2 | 78 | $0.003 | B+ |
| 11 | K Kimi K2.5Moonshot | 84.9 | 92 | $0.002 | B |
| 12 | G GPT-4o MiniOpenAI | 82.1 | 110 | $0.0006 | B |
| 13 | M Mistral Small 3Mistral · Open | 79.3 | 95 | $0.0004 | B |
| 14 | L Llama 3.1 8BMeta · Open | 76.4 | 145 | $0.0002 | C+ |
| 15 | P Phi-3.5Microsoft · Open | 73.8 | 130 | $0.0003 | C+ |
Most preferred models
VIBE Score (higher is better)
Fastest models in top 10
Output tokens per second
Best under $1 per 1M tokens
VIBE Score among low-cost models
Best open-source models
VIBE Score · open weights only