TestsBenchmarks · Human preference
LMArena Text - Hard Prompts
Crowd-voted, style-controlled Elo on prompts classified as hard (specific, complex, problem-solving).
Elo ratingHigher is betterCounts toward:Weights: Research & analysis ×0.5The test's websiteOfficial page ↗
Measured by independent testersIndependent Reported by the makerVendor-reported
| Thinking levelSetting | |||||||
|---|---|---|---|---|---|---|---|
| 1 | Gemini 4 Argon | Google (Gemini / DeepMind) | 1550 | Highgemini-4-argon-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 2 | Claude Opus 4.6 | Anthropic | 1533 | Highclaude-opus-4-6-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 3 | Claude Opus 5.5 | Anthropic | 1533 | Highclaude-opus-5.5-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 4 | Claude Fable 5 | Anthropic | 1531 | Highclaude-fable-5-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 5 | Claude Opus 4.7 | Anthropic | 1526 | Highclaude-opus-4-7-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 6 | Claude Fable 5.1 | Anthropic | 1522 | Maxclaude-fable-5.1-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 7 | Kimi K3 | Moonshot AI (Kimi) | 1518 | Maxkimi-k3-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 8 | Muse Spark 1.3 | Meta (Meta Superintelligence Labs, Muse) | 1517 | Maxmuse-spark-1.3-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 9 | Claude Opus 5 | Anthropic | 1516 | Highclaude-opus-5-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 10 | Gemini 3.8 Flash | Google (Gemini / DeepMind) | 1515 | Highgemini-3.8-flash-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 11 | Claude Opus 4.8 | Anthropic | 1513 | Highclaude-opus-4-8-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 12 | Muse Spark 1.1 | Meta (Meta Superintelligence Labs, Muse) | 1511 | Defaultmuse-spark-1.1 | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 13 | GPT-5.6 Sol | OpenAI | 1511 | Extra highgpt-5.6-sol-xhigh | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 14 | Gemini 3.7 Flash | Google (Gemini / DeepMind) | 1509 | Highgemini-3.7-flash-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 15 | Muse Spark 1.2 | Meta (Meta Superintelligence Labs, Muse) | 1508 | Extra highmuse-spark-1.2 (xHigh) | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 16 | Gemini 3.1 Pro (Preview) | Google (Gemini / DeepMind) | 1507 | Defaultgemini-3.1-pro-preview | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 17 | Muse Spark | Meta (Meta Superintelligence Labs, Muse) | 1506 | Defaultmuse-spark | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 18 | GLM-5.3 | Z.ai (Zhipu) | 1505 | Maxglm-5.3-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 19 | MiMo-V2.6-Pro | Xiaomi MiMo | 1505 | Defaultmimo-v2.6-pro | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 20 | Qwen3.8 Max (0803) | Qwen (Alibaba) | 1502 | Defaultqwen3.8-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 21 | GPT-6 Astra | OpenAI | 1501 | Maxgpt-6-astra-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 22 | DeepSeek V4.1 Flash | DeepSeek | 1498 | Maxdeepseek-v4.1-flash-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 23 | Claude Sonnet 5 | Anthropic | 1491 | Highclaude-sonnet-5-high | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 24 | GPT-6 Sol | OpenAI | 1484 | Maxgpt-6-sol-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 25 | DeepSeek V4 Pro (0813) | DeepSeek | 1483 | Highdeepseek-v4-pro-high-20260813 | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 26 | GPT-6 Luna | OpenAI | 1469 | Maxgpt-6-luna-max | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |
| 27 | Grok 4.7 | SpaceXAI (formerly xAI) | 1462 | Extra highgrok-4.7-xhigh | Independent testIndependentLMArena ↗ | — | 30 Sep 2026 |