- DeepSeek officially released its V4-Pro model on Thursday, making it available across the company's app, web interface, and API after the model had been in preview since April.
- The V4-Pro model focuses on agent capabilities, allowing AI systems to use tools, execute code, and complete multi-step workflows without human intervention.
- Benchmark results showed the V4-Pro model scored 87.9 on Terminal Bench 2.1, 62.7 on DeepSWE, and 61.5 on NL2Repo, among other agent-focused tests.
- DeepSeek's cheaper V4-Flash model performed above expectations in independent tests conducted after its release earlier this month, in some cases outpacing the April preview of V4-Pro.
- DeepSeek raised about $7.4 billion in its first round of outside capital in June 2026, marking a significant change for a company that had long avoided external financing.
- Reuters reported that DeepSeek was in discussions for an additional fundraising round at a valuation of about $74 billion.
- DeepSeek's R1 model went viral in early 2025, giving it a commanding position in the AI race, but competitors have since closed the gap.
- DeepSeek aims to regain ground against fast-moving domestic rivals as it expands hiring, computing capacity, and fundraising efforts.
- DeepSeek has said it aims to at least double staffing across departments, including data-centre and AI-agent teams.
- DeepSeek's fundraising reflects the growing cost of competing in AI, which requires large investments in computer chips, data centres, and specialized staff.
DeepSeek officially launched its V4-Pro AI model on August 13, 2026, after a preview period that began in April. The model, designated DeepSeek-V4-Pro-0813, enhances agent capabilities, allowing AI systems to execute complex tasks autonomously.
Benchmark results revealed that V4-Pro scored 87.9 on Terminal Bench 2.1, 62.7 on DeepSWE, and 61.5 on NL2Repo, showcasing its advanced performance.3
The model supports a context window of up to 1 million tokens and can generate outputs as long as 384,000 tokens. It also features three thinking effort levels—low, high, and max—allowing developers to tailor computation intensity to task complexity.
The launch follows the unexpected success of DeepSeek's V4-Flash model, which outperformed V4-Pro in independent tests, raising questions about the rapid technological advancements made by the company.
DeepSeek is also expanding its workforce, aiming to double staffing across departments, including AI-agent teams, and is planning a new fundraising round at a valuation of about $74 billion. This comes after raising $7.4 billion in June, marking a shift towards external capital to support its growth in the competitive AI landscape.
The company is also increasing private hiring of chip-design engineers to develop its own AI chips, reducing reliance on suppliers like Nvidia and Huawei.
“The V4-Pro-0813 model scores 87.9 on Terminal Bench 2.1 and supports a 1-million-token context window, with peak-hour output pricing rising to $3.96 per million from $0.87. DeepSeek's $7.4 billion June funding round and reported $74 billion valuation talks underscore its push to close the gap with rivals like Alibaba and ByteDance.”
qz.com
