Synthszr Charts — die großen AI-Marken im Wettkampf ums Podium

ERNIE 5.1

Baidu · v5.1 · 7× · last seen Jul 03, 2026

Momentum

ERNIE 5.1 is an optimized version of Baidu's 2025 ERNIE 5.0 with 2.4 trillion parameters. The model reduces parameter size by two-thirds while retaining 98% of the original model's performance. ERNIE 5.1 ranks number 4 on Arena's Search Leaderboard and operates at 6% of the compute cost of comparable models.

Momentum trend

19.05.17.08.

Features

Key Benchmark (%)	LMArena Search Leaderboard: 1,223 points (4th globally, 1st in China); AIME26 with tool use: 99.6%
Context Window (Tokens)	128,000 tokens (third-party reported; not officially confirmed by Baidu)
License	Proprietary, hosted-only (no open source, no public weights)
Multimodality	Text-only (no image/audio input); multimodal tasks require separate ERNIE-VL models
Platform	ERNIE Bot website (ernie.baidu.com), Baidu AI Studio Playground, Qianfan API (OpenAI-compatible)
Price per 1M Tokens	$0.59 input / $2.65 output per 1M tokens (Qianfan API)
Release Date	May 9, 2026 (official release); Preview on April 30, 2026

More products in this category: Frontier LLMs

Sources (7)

Company Analysis: Baidu

As of 08/08/2026

HOLDSynthszr Vote

With current price in EUR unknown, a strictly data-driven stance favors HOLD pending the next earnings update (mid-August 2026 per consensus calendars) to validate whether Q1 2026’s +49% YoY “AI-powered Business” growth and positive profitability (RMB 3.4bn net income; ~11% net margin) are sustained, and whether Apollo Go utilization continues scaling (3.2m fully driverless rides in Q1 2026; >350k weekly peak). (ir.baidu.com) If the stock is trading at a material discount to consensus targets and forward P/E remains near ~20x (USD-based), the risk/reward can justify upgrading to BUY after confirmation; conversely, any evidence of AI monetization deceleration, robotaxi regulatory friction, or margin deterioration would support de-risking. (boursorama.com)

Summary

Baidu is a China-focused internet and AI platform with two primary profit engines: (1) Baidu Core online marketing (search/feed ads) and (2) AI-driven growth vectors including Baidu AI Cloud (Qianfan platform + model services), AI applications, and autonomous driving (Apollo Go robotaxi). Its core competency is full-stack AI (models, tooling, cloud infrastructure, and distribution via search and apps), plus a large domestic user base and advertiser relationships that can be monetized as AI-enhanced search and marketing products mature. In recent months, the most material disclosed operating datapoints have been AI and robotaxi traction. In Q1 2026, Baidu reported “Baidu Core AI-powered Business” revenue above RMB 13.6bn (+49% YoY), with AI Applications revenue RMB 2.5bn (roughly flat YoY). Profitability remained positive: net income attributable to Baidu was RMB 3.4bn with an 11% net margin, and diluted EPS was RMB 8.76 per ADS. R&D expense was RMB 4.4bn, down 22% QoQ, indicating near-term cost discipline while continuing to fund AI and autonomous driving. (ir.baidu.com) On autonomous driving, Baidu disclosed that Apollo Go delivered 3.2m fully driverless operational rides in Q1 2026, with weekly rides peaking above 350k in March—useful evidence of scaling utilization (a key unit-economics driver). (sec.gov) On generative AI, Baidu publicly released ERNIE 5.1 (May 2026) and continues to commercialize via Qianfan model pricing, supporting a clearer monetization pathway for enterprise inference workloads. (ernie.baidu.com) Valuation and Street stance (in EUR terms) should be interpreted as “EUR-equivalent” because Baidu’s primary listings are USD/HKD. Consensus data points available from market aggregators indicate a forward P/E around ~20x (USD-based) and a FactSet-derived analyst consensus skewed to Buy/Outperform with a median target price materially above the then-current price (USD-based). (barchart.com) Short- to medium-term outlook is driven by: (a) stabilization/gradual recovery in online marketing, (b) AI cloud/model monetization (pricing + enterprise adoption), and (c) Apollo Go scaling without regulatory interruptions or reliability incidents. The next major catalyst is the upcoming earnings release (mid-August 2026 per consensus calendars), which should update AI revenue growth, margins, and robotaxi operating metrics. (barchart.com)

Key Takeaways

AI monetization is becoming more measurable: Baidu reported “AI-powered Business” revenue above RMB 13.6bn in Q1 2026 (+49% YoY), indicating AI is no longer only a narrative driver but a reported growth line. (ir.baidu.com)
Profitability remains positive with disclosed net income RMB 3.4bn and ~11% net margin in Q1 2026, alongside a QoQ reduction in R&D expense (RMB 4.4bn, -22% QoQ). (ir.baidu.com)
Apollo Go is scaling utilization: 3.2m fully driverless operational rides in Q1 2026 and weekly rides peaking above 350k in March, which matters for cost per ride and path to segment profitability. (sec.gov)
Commercialization infrastructure is explicit: Qianfan model pricing and the ERNIE 5.1 release provide concrete levers for enterprise inference revenue (pricing, usage, and attach to cloud). (ernie.baidu.com)
Street positioning is constructive: FactSet-derived consensus on Boursorama shows a majority “Acheter/Renforcer” style recommendations and a median target price well above the then-current price (USD-based), implying positive expected return if execution holds. (boursorama.com)

Action Ideas

BUY

Data-driven long exposure for investors seeking AI + autonomous optionality with current profitability. The thesis rests on (1) reported +49% YoY growth in “AI-powered Business” revenue in Q1 2026, (2) continued positive net income and double-digit net margin, and (3) scaling Apollo Go utilization (3.2m fully driverless rides in Q1 2026; >350k weekly peak), which can improve unit economics if reliability and regulatory conditions remain supportive. Valuation support is indicated by forward P/E around ~20x (USD-based) in market data and a positive analyst consensus skew. Position sizing should reflect China ADR/HKD listing risks and earnings volatility.

Horizon: 12 mo.

HOLD

Maintain exposure through the next earnings cycle if already positioned, focusing on confirmation of (a) AI-powered revenue growth durability beyond Q1 2026, (b) margin trajectory given R&D normalization, and (c) Apollo Go scaling metrics (rides, cities, cost per ride proxies). With the next earnings date indicated in mid-August 2026 by consensus calendars, waiting for updated KPIs can reduce decision error when current price and EUR FX translation are unknown. This is appropriate if portfolio constraints limit adding risk before new disclosures.

Horizon: 3 mo.

SELL

Reduce or exit if the investment mandate cannot tolerate China policy/ADR risk or if the portfolio requires near-term cash-flow visibility. Despite positive Q1 profitability, Baidu’s medium-term upside relies on scaling AI and robotaxi economics; if an investor requires low-variance earnings and minimal regulatory exposure, the risk-adjusted fit may be poor. A sell is also defensible if forward P/E (~20x USD-based) is viewed as full relative to near-term EPS growth expectations shown in consensus estimates (e.g., FY2026 vs FY2027 step-up), and if one expects competitive pricing to compress AI margins.

Horizon: 6 mo.

Contrarian Insights

• Robotaxi traction is already at a scale that can matter financially sooner than many investors assume: 3.2m fully driverless rides in Q1 2026 and >350k weekly peak rides suggest utilization is moving beyond pilot optics toward repeatable operations, which can change how investors model the segment (from long-dated option to nearer-term scaling business). (sec.gov)
• Cost discipline may be underappreciated versus the common view that AI investment necessarily drives persistent margin compression: Baidu reported a 22% QoQ decline in R&D expense in Q1 2026 while still reporting strong AI-powered revenue growth, implying management has levers to balance growth and profitability. (ir.baidu.com)