# 模型坐标

> 中文 AI benchmark 信息聚合站。公开来源指标保持原样；少数中文专题提供协议透明的 AI 复评分，不创建全站综合评分。

## Core pages
- Homepage: https://next.models.latent.vip/
- Decision guides: https://next.models.latent.vip/guides/
- Beginner model questions: https://next.models.latent.vip/faq/
- Task-specific rankings: https://next.models.latent.vip/rankings/
- Model comparisons: https://next.models.latent.vip/compare/
- Claude Code model guide: https://next.models.latent.vip/guides/best-model-for-claude-code/
- Cursor model guide: https://next.models.latent.vip/guides/best-model-for-cursor/
- Chinese coding model ranking: https://next.models.latent.vip/rankings/chinese-coding-models/
- DeepSeek vs Claude coding: https://next.models.latent.vip/compare/deepseek-vs-claude-coding/
- AI novel writing model ranking: https://next.models.latent.vip/rankings/ai-novel-writing-models/
- Chinese LLM ranking: https://next.models.latent.vip/rankings/chinese-llm/
- Model comparison: https://next.models.latent.vip/models/
- Full model directory: https://next.models.latent.vip/models/directory/
- Country analysis: https://next.models.latent.vip/countries/
- Coding agents: https://next.models.latent.vip/coding-agents/
- Media leaderboards: https://next.models.latent.vip/media/
- Capability indices: https://next.models.latent.vip/capabilities/
- Chinese evaluations: https://next.models.latent.vip/chinese-evaluations/
- Fortune-telling yearly-detail evaluation: https://next.models.latent.vip/chinese-evaluations/fortune-telling/
- Evaluation views: https://next.models.latent.vip/evaluations/
- Cost and performance: https://next.models.latent.vip/economics/
- API providers: https://next.models.latent.vip/providers/
- Hardware: https://next.models.latent.vip/hardware/
- Methodology: https://next.models.latent.vip/methodology/
- Data catalog: https://next.models.latent.vip/data/
- Update archive: https://next.models.latent.vip/updates/
- Machine-readable snapshot: https://next.models.latent.vip/data/homepage.json
- Machine-readable model metrics: https://next.models.latent.vip/data/models.json
- Machine-readable changes: https://next.models.latent.vip/data/changes.json
- Machine-readable Chinese evaluations: https://next.models.latent.vip/data/chinese-evaluations.json
- Full LLM guidance: https://next.models.latent.vip/llms-full.txt

## Public perspectives
- 智能指数: https://next.models.latent.vip/#intelligence (source: https://artificialanalysis.ai/#intelligence)
- 编码 Agent: https://next.models.latent.vip/#coding-agents (source: https://artificialanalysis.ai/agents/coding-agents)
- 图像与视频: https://next.models.latent.vip/#media-leaderboards (source: https://artificialanalysis.ai/#media-leaderboards)
- 语音: https://next.models.latent.vip/#speech-leaderboards (source: https://artificialanalysis.ai/#speech-leaderboards)
- 能力指数: https://next.models.latent.vip/#capability-indices (source: https://artificialanalysis.ai/models/capabilities)
- 中文测评: https://next.models.latent.vip/#chinese-evaluations (source: https://github.com/k2009/Kline-star/blob/main/docs/ai-model-evals/tencentmaas-yearly-report/compare-full-model-set-2026.md)
- 智能拆分: https://next.models.latent.vip/#intelligence-breakdown (source: https://artificialanalysis.ai/#intelligence-breakdown)
- AA-Briefcase: https://next.models.latent.vip/#aa-briefcase (source: https://artificialanalysis.ai/evaluations/aa-briefcase)
- AA-Omniscience: https://next.models.latent.vip/#omniscience (source: https://artificialanalysis.ai/evaluations/omniscience)
- GDPval-AA v2: https://next.models.latent.vip/#gdpval (source: https://artificialanalysis.ai/evaluations/gdpval-aa)
- 开放性指数: https://next.models.latent.vip/#openness (source: https://artificialanalysis.ai/evaluations/artificial-analysis-openness-index)
- 回答长度: https://next.models.latent.vip/#output-tokens (source: https://artificialanalysis.ai/#output-tokens)
- 价格与成本: https://next.models.latent.vip/#price-and-cost (source: https://artificialanalysis.ai/#price-and-cost)
- 速度与延迟: https://next.models.latent.vip/#speed (source: https://artificialanalysis.ai/#speed)
- API Provider: https://next.models.latent.vip/#providers (source: https://artificialanalysis.ai/#providers)
- 硬件: https://next.models.latent.vip/#hardware (source: https://artificialanalysis.ai/benchmarks/hardware)

## Provenance
- Primary general benchmark source: https://artificialanalysis.ai/
- Chinese evaluation source: https://github.com/k2009/Kline-star/blob/main/docs/ai-model-evals/tencentmaas-yearly-report/compare-full-model-set-2026.md
- Fortune-telling content scores are site-generated, sample-specific AI review scores with judge runs in the machine dataset.
- Snapshot date: 2026-07-21
- Each module links to its original source. Different benchmarks and units must not be merged as directly comparable scores.