DeepSeek Will Release V4.1 in June With Visual Multimodal Capabilities

Source status

This is a reported market and product update, not an official DeepSeek release note. DeepSeek has not yet published a final V4.1 model card, API changelog, pricing page, or investor announcement for this round.

The useful news is therefore narrow: track the reported funding structure, the reported June V4.1 release window, and the specific product capabilities attached to V4.1. Do not treat the numbers below as closed financing terms until DeepSeek or the investors confirm them.

What the new report says

The WeChat source says DeepSeek is preparing its first major outside financing round while also planning a June V4.1 update. The key reported points are:

• A maximum fundraising target of RMB 50 billion, or about $7.35 billion.
• A possible post-round valuation around RMB 350 billion, or about $51.5 billion.
• Founder Liang Wenfeng may contribute as much as RMB 20 billion, roughly 40% of the round described by the source.
• The China Integrated Circuit Industry Investment Fund, often called the Big Fund, is reported as a major state-backed participant. Some outside coverage describes the state-backed fund as leading the talks, while the WeChat source describes it as a possible second-largest investor.
• Alibaba's separate financing talks with DeepSeek are reported to have stalled, mainly around control, ecosystem binding, and DeepSeek's desire to remain independent.

Separate English coverage from TechCrunch, summarizing Financial Times and Bloomberg reporting, frames the same funding cycle as a potential first outside investment round that could value DeepSeek at roughly $45 billion. The valuation range is the cleaner takeaway than a single exact dollar figure.

V4.1 release window

The product side of the report says DeepSeek plans to release V4.1 in June 2026. The described update is not just a benchmark refresh. It is positioned around enterprise usability and tool-driven workflows:

• Stronger compatibility with MCP.
• More enterprise-grade tooling.
• Image and audio processing capability.
• Text remains the output format, according to the report.

For DeepSeek users, the most important point is that V4.1 may become the bridge between the current V4 text/coding family and a more practical multimodal workflow. Image and audio input would move DeepSeek closer to day-to-day screenshot analysis, document review, meeting/audio summarization, and visual QA workflows, while MCP support would matter for agent tools and developer automation.

Why MCP support matters

MCP is not a decorative feature. If DeepSeek V4.1 improves MCP compatibility, it could make DeepSeek easier to wire into coding agents, local developer tools, document pipelines, and controlled enterprise workflows.

That is especially relevant for this site because DeepSeek TUI, local development, and Claude Code-style integrations already depend on reliable tool surfaces. Better MCP behavior would reduce adapter friction and make DeepSeek easier to use as the primary model in agent workflows.

Why visual multimodality matters

The reported image and audio capability fits the broader direction already visible in DeepSeek's vision work. A text-only coding model can help with code, logs, and prompts. A visual multimodal model can also reason over screenshots, UI states, charts, diagrams, scanned documents, and image-backed support cases.

The report says V4.1 output will still be text. That limitation is important: this should be read as multimodal understanding, not image or audio generation.

Business context

The funding story explains why V4.1 timing matters. DeepSeek has built a strong research reputation, but the market is now asking whether it can move faster on products, retain senior researchers, and turn traffic into durable revenue.

The reported financing would give DeepSeek more room to compete for talent and infrastructure without becoming tightly bound to a single internet platform. That independence is part of the story: DeepSeek wants enough capital to commercialize faster while preserving its model-lab identity.

Editorial takeaway

The headline should not be inflated into "DeepSeek officially confirmed every V4.1 feature." The better reading is:

• Funding: first major outside round is being reported, with a valuation range around $45B-$51.5B depending on source framing.
• State capital: the Big Fund is reported as an important participant, but source wording differs on whether it leads or ranks second.
• Product: V4.1 is reportedly planned for June 2026, with MCP, enterprise tools, and visual/audio input as the important capability signals.
• User impact: the update would make DeepSeek more practical for agent workflows and multimodal understanding, while still keeping text output as the expected response format.

Sources checked

• WeChat source shared by the editor
• TechCrunch funding summary, May 6, 2026