Skip to main content

Tencent Cloud Phases Out Older AI Models: What Users Need to Know

Tencent Cloud Retires Legacy AI Models

In a move reflecting the rapid evolution of artificial intelligence technology, Tencent Cloud announced today it will sunset two older AI models later this month. The DeepSeek-V3 and DeepSeek-R1 models will officially go offline at midnight Beijing Time on November 24, 2025.

Image

Why the Change Matters

The tech giant cites continuous improvements in its large language model technology as the driving force behind this transition. "As our AI capabilities advance, we need to focus resources on supporting our most current and capable models," explained a Tencent Cloud spokesperson.

For businesses and developers relying on these models, the clock is ticking. After November 24, access to both discontinued versions will cease completely.

Smooth Transition Options

Tencent isn't leaving users stranded. The company recommends migrating to:

  • DeepSeek-V3-0324 (replacement for V3)
  • DeepSeek-R1-0528 (replacement for R1)

"We strongly urge all users to complete their migration before the cutoff date," the announcement stressed. "This ensures uninterrupted service and access to our latest improvements."

For those who miss the manual migration window, there's a safety net. Tencent will automatically shift remaining users to the new model versions—a move that should prevent any service disruptions but might require some adjustment period.

What This Means for Your Operations

The change highlights how quickly cloud-based AI services evolve. Developers using these models should:

  1. Test applications with the new versions immediately
  2. Update any documentation referencing old model names
  3. Budget time for potential integration tweaks

The automatic upgrade feature demonstrates Tencent's commitment to backward compatibility—but manual migration remains the preferred path for maintaining control over your AI implementations.

Key Points:

  • Deadline: November 24, 2025 (Beijing Time)
  • Affected Models: DeepSeek-V3 and DeepSeek-R1
  • Replacements: V3-0324 and R1-0528 respectively
  • Action Required: Migrate manually or prepare for automatic upgrade

Enjoyed this article?

Subscribe to our newsletter for the latest AI news, product reviews, and project recommendations delivered to your inbox weekly.

Weekly digestFree foreverUnsubscribe anytime

Related Articles

DeepSeek-V4 Set to Revolutionize Code Generation This February
News

DeepSeek-V4 Set to Revolutionize Code Generation This February

DeepSeek is gearing up to launch its powerful new AI model, DeepSeek-V4, around Chinese New Year. The update promises major leaps in code generation and handling complex programming tasks, potentially outperforming competitors like Claude and GPT series. Developers can expect more organized responses and better reasoning capabilities from this innovative tool.

January 12, 2026
AI DevelopmentProgramming ToolsMachine Learning
News

Snowflake Bets Big on Observe With $1B Deal to Power AI Monitoring

Snowflake makes its largest acquisition yet, snapping up observability platform Observe for $1 billion. The move strengthens Snowflake's position in AI-era data management by combining monitoring with its cloud data platform. Founded by Snowflake alumni and backed by the same VC firm, Observe brings native integration that promises 10x faster issue detection for customers wrestling with exploding AI system data.

January 9, 2026
SnowflakeAI MonitoringEnterprise Tech
News

Lenovo Unveils Qira AI Assistant and Teams Up With NVIDIA for Supercharged Cloud Computing

At CES 2026, Lenovo made waves by introducing Qira, its new personal AI assistant designed to evolve into users' digital twins. The tech giant also announced a groundbreaking partnership with NVIDIA to build massive AI cloud infrastructure capable of handling trillion-parameter models. These moves signal Lenovo's ambitious push into hybrid AI solutions that blend personal, enterprise and public intelligence.

January 7, 2026
AI AssistantsCloud ComputingHybrid Intelligence
News

DeepSeek Finds Smarter AI Doesn't Need Bigger Brains

DeepSeek's latest research reveals a breakthrough in AI development - optimizing neural network architecture can boost reasoning abilities more effectively than simply scaling up model size. Their innovative 'Manifold-Constrained Hyper-Connections' approach improved complex reasoning accuracy by over 7% while adding minimal training costs, challenging the industry's obsession with ever-larger models.

January 4, 2026
AI ResearchMachine LearningNeural Networks
Chinese AI Model Stuns Tech World with Consumer GPU Performance
News

Chinese AI Model Stuns Tech World with Consumer GPU Performance

Jiukun Investment's new IQuest-Coder-V1 series is turning heads in the AI community. This powerful code-generation model, running on a single consumer-grade GPU, outperforms industry giants like Claude and GPT-5.2 in coding tasks. Its unique 'code flow' training approach mimics real-world development processes, offering developers unprecedented creative possibilities while keeping hardware requirements surprisingly accessible.

January 4, 2026
AI DevelopmentMachine LearningCode Generation
News

Tencent Cloud's AI Revolution in Finance: 100+ Real-World Applications Now Live

Tencent Cloud has taken financial AI from theory to practice, deploying over 100 large model applications with major Chinese institutions. From detecting fraud in seconds to generating investment insights, these real-world implementations are transforming how banks and exchanges operate. The company emphasizes security and compliance as key differentiators in this sensitive sector.

December 31, 2025
Financial TechnologyAI ApplicationsTencent Cloud