As of 2026-09-13
Arena Text.
Their board. Their ranks.
Human preference Elo from the Arena.ai Text leaderboard.
Arena #1 is Fable 5
Arena Text
Arena.ai Text leaderboard. Preliminary rows are labelled. No averaging.
| 1 | Fable 5 Anthropic · closed New Claude 5 frontier tier | 1506Lab | #1Lab |
| 2 | Opus 4.7 Anthropic · closed Frontier coding/reasoning | 1502Lab | #3Lab |
| 3 | Spark 1.2 Meta · closed Agent-native Meta model | 1500Lab | #4Lab |
| 4 | Fable 5.1 Anthropic · closed New Claude 5 frontier tier | 1498Lab | #5Lab |
| 5 | 3.8 Flash Google DeepMind · closed Latest Flash workhorse | 1493Prelim | #9Prelim |
| 6 | 3.7 Flash Google DeepMind · closed Latest Flash workhorse | 1490Prelim | #12Prelim |
| 7 | Opus 5 Anthropic · closed Anthropic’s flagship frontier model | 1487Lab | #14Lab |
| 7 | 3.1 Pro Google DeepMind · closed Frontier multimodal/reasoning | 1487Lab | #15Lab |
| 9 | Kimi K3 Moonshot AI · open Frontier open-weight | 1485Lab | #17Lab |
| 10 | GPT-5.6 Sol OpenAI · closed Top-tier flagship reasoning | 1483Lab | #18Lab |
| 10 | GLM-5.3 Z.ai · staged Latest GLM coding / cyber-defense | 1483Lab | #19Lab |
| 12 | GPT-5.5 OpenAI · closed Frontier reasoning + agents | 1482Lab | #20Lab |
| 13 | Qwen3.8 Max Alibaba · closed Hosted 2.4T flagship; weights promised | 1481Lab | #22Lab |
| 13 | Opus 4.8 Anthropic · closed Frontier reasoning/coding | 1481Lab | #21Lab |
| 15 | 3.6 Flash Google DeepMind · closed Frontier-class fast reasoning | 1480Lab | #23Lab |
| 16 | Grok 4.20 SpaceXAI · closed Frontier reasoning/agents | 1475Lab | #30Lab |
| 17 | Qwen3.7 Max Alibaba · closed Frontier Chinese model | 1473Lab | #34Lab |
| 18 | GLM-5.2 Zhipu AI · open Frontier/open-weight coding | 1472Lab | #38Lab |
| 19 | Grok 4.5 SpaceXAI · closed Frontier coding/agents | 1468Lab | #42Lab |
| 20 | GPT-5.6 Terra OpenAI · closed Balanced GPT-5.6 | 1466Lab | #45Lab |
| 21 | Sonnet 5 Anthropic · closed Frontier general-purpose/agentic | 1461Lab | #51Lab |
| 22 | Grok 4.6 SpaceXAI · closed Latest Grok frontier model | 1456Lab | #63Lab |
| 23 | GPT-5.6 Luna OpenAI · closed Cost-efficient GPT-5.6 | 1452Lab | #67Lab |
| 24 | Grok 4.3 SpaceXAI · closed Frontier reasoning/agents | 1443Lab | #81Lab |
24 models on this board · # is this page’s rank