Updated May 2, 2026

AI Power Rankings

IndependentΒ·Frontier AI lab index
← Companies
πŸ‡¨πŸ‡³

Baidu

Beijing, China

ERNIE 5.0 (2.4T params); 200M+ monthly AI assistant users.

Efforts: ERNIE 5.0, Ernie Bot, Baidu Cloud AI, custom AI chips
Rank
#9
Structural
7.25
Unweighted
7.16
vs #1
-2.21

Profile across ten areas

Area breakdown

🧠Model Quality
Core Engine Β· 18%
ERNIE 5.0 (2.4T parameters, sparse MoE) and ERNIE 5.1 Preview (April 29, 2026, score 1,476 on LMArena Text Arena) make Baidu the #1 Chinese model and #13 globally, also topping the Legal AI leaderboard; still does not compete at the very frontier with GPT-5.5 or Claude Opus 4.7 on broad international benchmarks.
β€” Claude Opus 4.7
Claude: 7.4Claude: 7.2Gemini: 7.2GPT-5.5: 7.5
7.3
panel mean
πŸ”¬Research
Core Engine Β· 12%
Custom AI chip development (Kunlunxin M100/M300), autonomous driving research (Apollo Go), and the ultra-large sparse MoE architecture show technical breadth; but fundamental AI research publications have declined, and Baidu is increasingly outpaced by DeepSeek and Alibaba in model innovation.
β€” Claude Opus 4.6
Claude: 6.3Claude: 6.3Gemini: 7.4GPT-5.5: 7.3
6.8
panel mean
πŸ—οΈProduct
Delivery Β· 10%
ERNIE Assistant, Baidu Search integration, Baidu Cloud AI, and app partnerships with JD, Meituan, and Trip.com provide a broad China-focused platform; the limitation is weaker global developer and consumer mindshare than leading international platforms.
β€” GPT-5.5 Pro
Claude: 7.2Claude: 7.2Gemini: 8.3GPT-5.5: 7.7
7.6
panel mean
πŸ’°Business
Accelerants & Stabilizers Β· 5%
Baidu is a major Chinese tech company with strong cloud and AI revenue, and Apollo Go's 17M rides prove real-world AI monetization; but core search revenue is under pressure from AI chatbots and ByteDance, and the company's growth trajectory lags newer Chinese AI companies.
β€” Claude Opus 4.6
Claude: 6.4Claude: 6.4Gemini: 8.1GPT-5.5: 7.4
7.1
panel mean
⚑Compute
Core Engine Β· 14%
Custom Kunlunxin AI chips (M100 for inference in 2026, M300 for training in 2027) demonstrate serious infrastructure investment and domestic chip capability; but chip performance likely trails NVIDIA and Google TPUs, and total data center capacity is smaller than US hyperscalers.
β€” Claude Opus 4.6
Claude: 6.8Claude: 6.8Gemini: 8.6GPT-5.5: 7.8
7.5
panel mean
πŸ›‘οΈSafety
Accelerants & Stabilizers Β· 6%
Baidu operates under Chinese AI regulatory controls and has enterprise cloud governance, but public model-card, red-team, and alignment disclosures are thinner than Western frontier labs; the counterpoint is that search-scale deployment forces operational moderation.
β€” GPT-5.5 Pro
Claude: 4.3Claude: 4.3Gemini: 7.6GPT-5.5: 6.8
5.8
panel mean
πŸ‘₯Talent
Delivery Β· 10%
Established AI team with deep experience in search, NLP, and autonomous driving; but top AI talent is being poached by DeepSeek, ByteDance, and Moonshot, and Baidu's legacy organizational culture can slow AI innovation.
β€” Claude Opus 4.7
Claude: 5.9Claude: 5.9Gemini: 7.7GPT-5.5: 7.4
6.7
panel mean
🌐Distribution
Delivery Β· 8%
Baidu can distribute ERNIE through search, cloud, maps, mobile apps, and Ernie Assistant, which is reported at 200M MAU; the limitation is that its reach is more China-centric and less global than Google, Meta, Apple, Microsoft, Amazon, or ByteDance.
β€” GPT-5.5 Pro
Claude: 7.1Claude: 7.1Gemini: 8.7GPT-5.5: 8.3
7.8
panel mean
🏰Data & Moats
Core Engine Β· 12%
Chinese search index data, Maps data, Apollo Go driving data (17M+ rides), and the Ernie Bot interaction flywheel create substantial domestic moats; these advantages are geographically limited and the search moat is eroding as users shift to AI-native interfaces.
β€” Claude Opus 4.7
Claude: 7.1Claude: 7.1Gemini: 9.3GPT-5.5: 8.5
8.0
panel mean
πŸš€Momentum
Accelerants & Stabilizers Β· 5%
ERNIE 5.1 Preview just took the #1-Chinese-model crown back from DeepSeek-V4-Pro within days, topped the global Legal AI leaderboard, and the full launch is teed up for Baidu Create 2026; counterweight is that core search is still under structural pressure and Baidu is no longer the trendsetter in Chinese AI.
β€” Claude Opus 4.7
Claude: 6.4Gemini: 6.8GPT-5.5: 7.7
7.0
panel mean

News log

Baidu β€” News Log

Tracked for AI Power Rankings scoring. Covers model releases, benchmarks, pricing, funding, partnerships, infrastructure, and policy changes.


2026-01-15 β€” Baidu's ERNIE 5.0 ranks first among Chinese models; custom AI chips announced

Published: 2026-01-15 | Logged: 2026-05-02T09:00Z | Area: Model Quality, Compute & Infra, Product & Platform

Baidu's ERNIE 5.0, a 2.4 trillion parameter omni-modal model (text, images, audio, video), ranked first among Chinese models and eighth globally on LMArena text benchmarks, ahead of GPT-5.1-High and Gemini-2.5-Pro. The model uses an ultra-large MoE architecture with less than 3% of parameters active per inference. Baidu also announced two custom AI chips: Kunlunxin M100 (2026) and M300 (2027). The Ernie Bot assistant has reached 200 million monthly active users, and Baidu's 2026 strategy integrates AI directly into search rather than as a separate product.

Scoring impact: Positive for Model Quality (top Chinese model ranking). Compute & Infra boosted by custom chip development. Product & Platform strengthened by 200M MAU and search integration strategy.

Sources:


2026-04-29 β€” ERNIE 5.1 Preview launches as #1 Chinese model on LMArena

Published: 2026-04-29 | Logged: 2026-05-02T09:00Z | Area: Model Quality, Business & Market

Baidu released ERNIE 5.1 Preview, which shot to #13 globally on LMArena's Text Arena (score 1,476), becoming #1 Chinese AI model on the leaderboard. Beat DeepSeek-V4-Pro which launched just days earlier. ERNIE 5.1 Preview is a lighter, cheaper version atop the 5.0 foundation. Also topped the Legal AI leaderboard worldwide. Full launch confirmed for Baidu Create 2026 developer conference.

Scoring impact: Strong Model Quality boost (#1 Chinese model, global top 15). Positive for Business & Market (competitive positioning, efficiency gains).

Sources:


← Higher rank
πŸ‡ΊπŸ‡Έ Amazon
#8 Β· 7.77
Lower rank β†’
πŸ‡ΊπŸ‡Έ xAI
#10 Β· 7.23