Google DeepMind executive claims widening gap between closed and open AI models
Melvin Johnson, vice president of research at Google DeepMind, stated that closed frontier labs maintain a growing lead over open-weight artificial intelligence models.
FTMQ SI, written by our newsroom0 views
Closed frontier laboratories are widening their lead over open-weight models, according to statements made by Melvin Johnson, vice president of research at Google DeepMind. Finance.biggo.com and BigGo Finance both reported that Johnson described the performance gap between closed proprietary systems and open models as wider than at any previous point in 2026. [1][2]
Finance.biggo.com reported that Johnson, who leads Gemini post-training at DeepMind, outlined structural advantages supporting closed labs. These advantages include proprietary Tensor Processing Units, internal application testing, and direct distribution channels. Johnson also presented Gemini 4 Hydrogen, detailing its benchmark performance metrics, execution speed, cost structure, enterprise deployment cases, and planned agentic features for the Gemini application. [1]
As FTMQ SI reported earlier, DeepMind product-focused employees recently applied to the United Kingdom Central Arbitration Committee for a formal union recognition ballot. That application was filed by the Communication Workers Union and Unite the Union. Additionally, earlier testing by evaluation firm Vals showed open-weight Chinese models matching closed proprietary architectures in simulated multi-step cyber operations. [5][6]
OpenAI operates as an American public benefit corporation in San Francisco that develops generative artificial intelligence systems, as documented by Wikipedia. Finance.biggo.com reported that Google DeepMind intends to leverage its hardware and distribution infrastructure to deploy agentic capabilities across its product ecosystem. [1][4]
In short
- Melvin Johnson stated that the performance gap between closed and open AI models is at its widest point in 2026.
- Google DeepMind cited TPUs and internal testing as key structural advantages for closed labs.
- The remarks were delivered alongside a product briefing on the Gemini 4 Hydrogen model.
Sources
Every paragraph above points to the numbered items it rests on. Read the originals here.
- [1]Closed Frontier Labs Are Pulling Ahead of Open AI | Melvin Johnson|MTSfinance.biggo.com, 3d ago (the report this story comes from)
- [2]Google DeepMind's Melvin Johnson Says the Closed-Open AI Gap Is the Widest It Has Been All YearBigGo Finance, 4d ago
Background
- [3]Peter Thiel on Wikipedia
- [4]OpenAI on Wikipedia
- [5]Google DeepMind staff request UK arbitration body to trigger union ballot FTMQ SI, 4d ago
- [6]Open Chinese AI models approach closed frontier capabilities in network hacking tests FTMQ SI, 11d ago
Our newsroom writes these reports with the help of software, from the 6 sources listed and nothing else, and checks them against those sources. Facts can still be wrong or move on; the originals are the record. Spotted a mistake? Write to daniel@monsterkong.com.
Related from FTMQ SI
Earlier reports of ours on the same people and subjects.
- Google DeepMind staff request UK arbitration body to trigger union ballotIndustry Labs, 4d ago
- Isomorphic Labs raises 5 billion dollars to advance AI drug designIndustry Labs, 4h ago
- Open Chinese AI models approach closed frontier capabilities in network hacking testsTop Stories, 11d ago
- United States Artificial Intelligence Leaders Suggest Slowing Down Frontier DevelopmentIndustry Labs, 35d ago
More in Industry Labs
- Isomorphic Labs raises 5 billion dollars to advance AI drug design4h ago
- OpenAI releases GPT-6.2 Astra and Sol models4h ago
- Anthropic moves to float as major tech firms hold key stakes4h ago
- OpenAI moves forward with AMD partnership to diversify compute infrastructure4h ago
- Microsoft Research releases TokenEconomizer plugin to lower coding agent token usage1d ago
- Demonstrators rally in San Francisco demanding international treaty on advanced AI2d ago
Get the day in one email
Reports like this one, the top news of the last 24 hours, every morning. Free, one email a day; readers can comment under every report.
By signing up you agree to our terms and privacy policy. Unsubscribe any time.

Comments
Loading
Join the conversation
Comments are open to readers of our daily email: the top news of the last 24 hours, every morning, free. Sign up and the comment box opens.
Already on the list? Enter the same address and we will send a sign-in link.
By signing up you agree to our terms and privacy policy. Unsubscribe any time.