Model Leaderboard
🚀 View Latest Releases →
1314 models
Vendor
Open Source
Price Range
Domain
About the composite score
- Composite = Human Preference percentile × 60% + Intelligence Index percentile × 40% (weights adjustable)
- Human Preference: best of LM Arena text / vision / WebDev boards; requires ≥2 boards with data
- Intelligence Index: aggregated from third-party benchmarks such as Artificial Analysis
- Percentile normalization: each source is converted to a 0–100 percentile within the current model pool, reflecting relative position
- Fallback: no preference data → intelligence only; no intelligence → preference only; neither → "—"
- Last updated: 08/30
| Rank | Compare | Model | Score | Price | Context |
|---|---|---|---|---|---|
| 1 | 100 | $25 | 1M | ||
| 2 | 99.5 | — | — | ||
| 3 | 99.5 | $50 | 1M | ||
| 4 | 99.3 | $10 | 1.1M | ||
| 5 | 98.9 | $15 | 1.0M | ||
| 6 | 98.7 | $4.4 | 1.0M | ||
| 7 | 98.7 | $6 | 1M | ||
| 8 | glm-5.3-max Try →
| 98.6 | — | — | |
| 9 | gemini-3.7-flash-high Try →
| 98.4 | — | — | |
| 10 | glm-5.2-max Try →
| 98.2 | — | — | |
| 11 | 98 | $6 | 1.0M | ||
| 12 | 97.9 | — | — | ||
| 13 | 97.5 | — | — | ||
| 14 | 97.2 | $12 | 1.1M | ||
| 15 | 97 | $6 | 500K | ||
| 16 | Qwen3.8-Flash-Next Try →
| 96.6 | — | — | |
| 17 | 96.3 | — | — | ||
| 18 | 96.2 | $10 | 1M | ||
| 19 | gemini-3.6-flash-high Try →
| 96.1 | — | — | |
| 20 | 95.9 | $2.55 | 1M |