Live ranking, refreshed every 2 days
Best AI Models for German
As of October 8, 2026, the top-ranked model on LMArena's text leaderboard for German prompts is gemini-3-pro by Google with a score of 1,521, 3 points ahead of muse-spark-1.3-max by Meta.
Updated October 13, 2026. Showing the last saved copy of the leaderboard, dated October 8, 2026, because the live source could not be reached.
Ranking: top 10 models
Scores come from blind head-to-head votes on LMArena's text leaderboard for German prompts. Rank is LMArena's published rank, which accounts for confidence intervals, so several models can share a rank. The top 10 are shown below, and ranks 11 to 40 follow in the expandable list.
| Rank | Model | Lab | Score | Votes | License |
|---|---|---|---|---|---|
| 1 | gemini-3-pro | 1,521 ±22 | 813 | Proprietary | |
| 2 | muse-spark-1.3-max | Meta | 1,518 ±39 | 219 | Proprietary |
| 3 | glm-5.3-max | Z.ai (Zhipu) | 1,514 ±31 | 330 | Open weights |
| 4 | muse-spark | Meta | 1,511 ±40 | 237 | Proprietary |
| 5 | deepseek-v4.1-flash-max | DeepSeek | 1,509 ±40 | 191 | Open weights |
| 6 | claude-opus-4-6-high | Anthropic | 1,507 ±17 | 1,274 | Proprietary |
| 7 | gpt-5.6-sol-xhigh | OpenAI | 1,505 ±24 | 609 | Proprietary |
| 8 | claude-opus-5-max | Anthropic | 1,505 ±27 | 513 | Proprietary |
| 9 | claude-opus-4-7-high | Anthropic | 1,504 ±19 | 1,022 | Proprietary |
| 10 | kimi-k3-max | Moonshot AI | 1,502 ±25 | 599 | Open weights |
Show ranks 11 to 40
Rank Model Lab Score License 11 claude-fable-5-high Anthropic 1,501 ±24 Proprietary 12 gemini-3-flash Google 1,499 ±25 Proprietary 13 claude-opus-4-7 Anthropic 1,498 ±19 Proprietary 14 gemini-3.7-flash-high Google 1,494 ±31 Proprietary 15 gemini-3.1-pro-preview Google 1,494 ±15 Proprietary 16 claude-opus-4-8-high Anthropic 1,492 ±19 Proprietary 17 claude-opus-4-6 Anthropic 1,491 ±17 Proprietary 18 grok-4.20-beta1 xAI 1,490 ±28 Proprietary 19 gpt-5.5 OpenAI 1,490 ±19 Proprietary 20 gemini-3.5-flash-medium Google 1,490 ±23 Proprietary 21 claude-fable-5.1-max Anthropic 1,489 ±37 Proprietary 22 claude-opus-5-high Anthropic 1,488 ±20 Proprietary 23 gemini-3.8-flash-high Google 1,487 ±26 Proprietary 24 qwen3.8-max Alibaba (Qwen) 1,487 ±29 Proprietary 25 gemini-3.5-flash-high Google 1,484 ±22 Proprietary 26 gpt-5.2-chat-latest-20260210 OpenAI 1,482 ±25 Proprietary 27 gemini-3.6-flash-high Google 1,481 ±25 Proprietary 28 claude-opus-4-5-20251101 Anthropic 1,478 ±18 Proprietary 29 gpt-5.6-terra-xhigh OpenAI 1,478 ±24 Proprietary 30 muse-spark-1.1 Meta 1,478 ±23 Proprietary 31 gpt-4.5-preview-2025-02-27 OpenAI 1,477 ±32 Proprietary 32 claude-opus-4-5-20251101-high-32k Anthropic 1,477 ±23 Proprietary 33 glm-5.3-flash Z.ai (Zhipu) 1,477 ±27 Open weights 34 gpt-5.4-high OpenAI 1,476 ±19 Proprietary 35 gemini-3-flash (thinking-minimal) Google 1,476 ±16 Proprietary 36 gpt-5.6-luna-xhigh OpenAI 1,476 ±23 Proprietary 37 glm-5.2-max Z.ai (Zhipu) 1,475 ±21 Open weights 38 grok-4.20-multi-agent-beta-0309 xAI 1,474 ±19 Proprietary 39 gpt-6-astra-max OpenAI 1,473 ±47 Proprietary 40 ernie-5.1 Baidu 1,472 ±22 Proprietary
Showing the top 40 of 312 models on this leaderboard. Data from October 8, 2026.
Best model by lab
Each lab's highest-ranked model on this leaderboard, and how far it sits from the top.
| Lab | Best model | Rank | Score | Behind #1 |
|---|---|---|---|---|
| gemini-3-pro | 1 | 1,521 | Leader | |
| Meta | muse-spark-1.3-max | 2 | 1,518 | 3 pts |
| Z.ai (Zhipu) | glm-5.3-max | 3 | 1,514 | 7 pts |
| DeepSeek | deepseek-v4.1-flash-max | 5 | 1,509 | 11 pts |
| Anthropic | claude-opus-4-6-high | 6 | 1,507 | 14 pts |
| OpenAI | gpt-5.6-sol-xhigh | 7 | 1,505 | 16 pts |
| Moonshot AI | kimi-k3-max | 10 | 1,502 | 19 pts |
| xAI | grok-4.20-beta1 | 18 | 1,490 | 31 pts |
Which AI lab leads?
The lab with the most models at the top of the board is the one with the deepest bench, not only the single best model.
Anthropic holds 3 of the top 10 models (30%).
Models in the top 10 by lab
Show data table
| Lab | Top 10 |
|---|---|
| Anthropic | 3 |
| Meta | 2 |
| DeepSeek | 1 |
| 1 | |
| Moonshot AI | 1 |
| OpenAI | 1 |
| Z.ai (Zhipu) | 1 |
Open vs proprietary: how big is the gap?
The best open-weight model is glm-5.3-max by Z.ai (Zhipu) at 1,514, 7 points behind the best proprietary model, gemini-3-pro by Google at 1,521. Their 95% confidence intervals overlap, so the data cannot separate them.
2 proprietary models rank above the best open-weight model, and 3 of the top 20 are open-weight.
Open-weight here means any license other than Proprietary, including licenses with usage limits. It is a broader definition than the OSI one.
Other leaderboards
LMArena runs separate arenas for different jobs. The best model for one is often not the best for another.
Other arenas
- Text (overall)
- WebDev: claude-opus-5.5-max by Anthropic, 1,813
- Vision: claude-fable-5-high by Anthropic, 1,308
- Search: gpt-5.6-sol-xhigh by OpenAI, 1,256
- Document: claude-fable-5 by Anthropic, 1,499
- Text to image: gpt-image-2.5-sunburst by OpenAI, 1,425
- Image editing: gpt-image-2.5-sunburst by OpenAI, 1,524
- Text to video: gemini-omni-1.1-flash by Google, 1,516
Text arena by category
How to read this leaderboard
This category filters the text arena to prompts written in German. It measures which model people prefer when they ask in that language, which can differ from the English ranking.
- Scores come from blind votes
- LMArena shows a person two anonymous answers to the same prompt and asks which is better. Those head-to-head votes are fitted with a Bradley-Terry model, the statistical successor to Elo, to give each model one arena score. On the Elo scale LMArena uses, a 100-point lead means the higher model is expected to win about 64% of matchups.
- Confidence intervals decide ties
- Every score carries a 95% confidence interval. When two intervals overlap, the order of those two models is not settled. LMArena's rank column already accounts for this, which is why more than one model can hold the same rank.
- Why it differs from usage share
- A high rank does not mean many people use a model. Usage depends on price, distribution and habit. For adoption rather than quality, see AI market share.
Methodology and source
All rankings are read from the LMArena leaderboard dataset on Hugging Face, published by LMArena (formerly Chatbot Arena) under the Creative Commons Attribution 4.0 license. Credit for the underlying votes, scores and confidence intervals belongs to LMArena. Essentras is not affiliated with LMArena.
This page reads the text_style_control subset, category german, from the dataset's latest split. Scores, intervals, ranks, vote counts and license labels are shown as published. Essentras adds the grouping by lab and the derived figures on this page: gaps, lab shares, open vs proprietary and, on the main page, the history of who held #1.
Data refreshes automatically every 2 days. If the source cannot be reached, the page falls back to the last saved copy and says so at the top. Models are never merged or renamed; model and lab names appear exactly as LMArena publishes them.
Open-weight means any license other than Proprietary. History is limited to what the dataset records: the style-controlled text leaderboard from August 28, 2024. We do not scrape any other ranking site.
Track how AI models rank and describe your brand with our AI SEO service. Want these models to name your brand? Read our guide to LLM SEO.
Frequently asked questions
What is the best AI model for German?
As of October 8, 2026, gemini-3-pro by Google ranks first on LMArena's text leaderboard for German prompts with a score of 1,521. Rankings move as models and votes are added, so check the table above for the current order.
Which AI lab leads for German?
Anthropic holds 3 of the top 10 models (30%), ahead of Meta (2), DeepSeek (1), and Google (1).
Is the best open-source model close to the best proprietary model?
The best open-weight model is glm-5.3-max by Z.ai (Zhipu) at 1,514, 7 points behind the best proprietary model, gemini-3-pro by Google at 1,521. 2 proprietary models rank above the best open-weight model, and 3 of the top 20 are open-weight. Their 95% confidence intervals overlap, so the data cannot separate them. "Open" here means any license other than Proprietary, which is broader than the OSI definition.
What does this leaderboard measure?
This category filters the text arena to prompts written in German. It measures which model people prefer when they ask in that language, which can differ from the English ranking.
How often does this leaderboard update?
This page re-reads LMArena's dataset every 2 days. LMArena publishes a new snapshot when it has gathered enough votes or added models, so the date at the top shows the latest published data.
Models rank themselves. What do they say about you?
A leaderboard shows which AI models are strongest. It does not show what those models tell your customers about your company. Run the free check and see the answers.