Higher Than GPT-5? We Strive ERNIE X1.1, Baidu’s Newest AI Mannequin

Take: The construction appears neat. Headings keep disciplined. Desk formatting holds.

Picture era: reasoning-guided format and variant management

• Aim: Design a 1080×1350 occasion poster, then create a clear variant.
• Why this issues: I anticipate ERNIE X1.1 to plan earlier than rendering. Reasoning ought to information instrument calls.

Immediate:

“Design a 1080×1350 poster for ‘Information + Dharma 2025’. Mix a skyline, a temple silhouette, and a circuit board sample. Place the hero title top-center. Put the subtitle bottom-left. Place sponsor logos bottom-right. Preserve distinction excessive and textual content readable on cell.”

Observe-up immediate:

“Create a second variant. Change subtitle to ‘Historical Concepts. Trendy Intelligence.’ Preserve format similar. Improve brand padding by 12 pixels.”

Output:

Take: ERNIE X1.1 clearly failed right here. It looks as if the mannequin couldn’t handle the nitty-gritties of the immediate, and took a very long time to provide you with a fully garbage output.

Code Era: FastAPI service for hallucination threat

• Aim: Ship a tiny service that scores hallucination threat. Embody unit checks.
• Why this issues: I need scaffolds that run on the primary strive. ERNIE X1.1 ought to ship a clear construction.

Immediate:

“Write a Python FastAPI service /threat that accepts {textual content, sources[]} and returns hallucination_risk in [0,1]. Use lexical overlap and a easy contradiction verify. Embody pytest unit checks. No exterior SaaS calls. Preserve it production-readable.”

Output:

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from typing import Record

app = FastAPI()

class RiskRequest(BaseModel):
    textual content: str
    sources: Record[str]

class RiskResponse(BaseModel):
    hallucination_risk: float

def calculate_lexical_overlap(textual content: str, sources: Record[str]) -> float:
    """Calculate Jaccard similarity between textual content and sources"""
    text_words = set(textual content.decrease().cut up())
    total_similarity = 0.0
    valid_sources = [s for s in sources if s.strip()]
    
    if not valid_sources:
        return 0.0
    
    for supply in valid_sources:
        source_words = set(supply.decrease().cut up())
        intersection = text_words & source_words
        union = text_words | source_words
        similarity = len(intersection) / len(union) if union else 0
        total_similarity += similarity
    
    return total_similarity / len(valid_sources)

def calculate_contradiction(textual content: str, sources: Record[str]) -> float:
    """Easy NLI-inspired contradiction detection"""
    if not sources:
        return 0.0
    
    negation_words = {'not', 'no', 'by no means', 'none', 'with out'}
    contradiction_count = 0
    
    for supply in sources:
        source_tokens = supply.decrease().cut up()
        text_tokens = textual content.decrease().cut up()
        
        # Examine for direct contradictions
        for i, token in enumerate(text_tokens):
            if token in negation_words and that i+1 

Early Impressions

Right here is my sincere take up to now – ERNIE X1.1 thinks so much. It second-guesses many steps. Easy duties typically set off lengthy inner reasoning, slowing easy outputs that you just anticipate to be fast.

On some prompts, ERNIE X1.1 feels overcautious. It insists on planning past the duty. The additional pondering typically hurts coherence. Quick solutions grow to be meandering and not sure, similar to a human overthinking.

When ERNIE X1.1 hits the groove, it behaves effectively. It respects format and part order, and might maintain tables tight and codes neat. The “assume time,” although, typically feels heavy.

In my future use of it, I’ll tune prompts to curb this by lowering instruction ambiguity and including stricter constraints. For on a regular basis drafts, the additional pondering wants restraint. ERNIE X1.1 exhibits promise, but it surely should tempo itself.

Limitations and Open Questions

Entry exterior China nonetheless includes friction on cell. ERNIE X1.1 works finest by way of the net or API interface. Pricing particulars stay unclear at launch. I additionally need exterior benchmark checks, as the seller claims on the time of launch sound too daring to be correct.

The “pondering” depth wants person management. A visual knob might assist on this regard. If it have been to me, I’d add a quick mode to the mannequin for all these fast drafts and emails. Then once more, a deep mode for brokers and instruments can be useful as effectively. ERNIE X1.1 can profit from clear distinctions.

Conclusion

ERNIE X1.1 goals for reliability, not flash. The declare is fewer hallucinations and higher compliance. My runs present sturdy construction and first rate code. But the mannequin typically overthinks. That hurts pace and coherence on easy asks.

I’ll maintain testing with tighter prompts. I’ll lean on API paths for brokers. If Baidu exposes “assume” management, adoption will rise. Till then, ERNIE X1.1 stays in my toolkit for strict drafts and clear scaffolds. It simply must breathe between ideas.