Skip to content

Debates

Will the US close its gap with China in open-weights models?

8 recorded positions from 4 people, first said Nov 17, 2025. They do not agree — the readings below are what each one actually argued.

Us behind but momentum is building

Anastasios Angelopoulos · Aug 3, 2026

Thinking Machines' result is impressive or unimpressive depending on framing: it is the number one American open-source model after only six months of work, but the company is a year and a half old and sits only tenth overall with nine Chinese models above it

Arena leaderboard data shows nine Chinese open-source models ranked above the top American open-source model

Scope: per Arena leaderboard data today; generous vs. less generous reading of the same result

13:55 20VC: 70% of Neolabs Will Die | There Will be a $100BN US Open-Source Model | Data is a Trillion $ Market | Governments Cannot Regulate Models: It is Too Late | The Cyber Attacks to Come Will be Insane with Anastasios Angelopoulos @ Arena

Alex Atallah · Aug 10, 2026

America is still very far behind in open weight models, and GLM 5.2 was a major step forward for open weight models

Scope: comparison is with Chinese open models

0:00 20VC: Will OpenRouter Sell for $10BN to Stripe? | Why Chinese Open Models Are Beating America—and What Happens Next | Why Enterprises Are More Fearful of Anthropic and OpenAI Than China | Is the Routing Layer Becoming a Commodity with Alex Atallah

Alex Atallah · Aug 10, 2026

America is very far behind on open models and we should be concerned about the rate and quality of Chinese open models

Though things are picking up with Poolside, Thinking Machines and RC

Scope: says the situation is improving

27:19 20VC: Will OpenRouter Sell for $10BN to Stripe? | Why Chinese Open Models Are Beating America—and What Happens Next | Why Enterprises Are More Fearful of Anthropic and OpenAI Than China | Is the Routing Layer Becoming a Commodity with Alex Atallah

China holds the open source lead and is accelerating

Andrew Ng · Nov 17, 2025

China has taken a lead in releasing high-quality open-weight models, which nobody would have predicted a decade ago

Scope: 'taking a lead or getting up there'

21:12 20VC: Andrew NG on The Biggest Bottlenecks in AI | How LLMs Can Be Used as a Geopolitical Weapon | Do Margins Matter in a World of AI? | Is Defensibility Dead in a World of AI? | Will AI Deliver Masa Son's Predictions of 5% GDP Growth?

Harry Stebbings · Jul 6, 2026

China has the best open source models and its open source ecosystem is evolving at an unbelievable rate

26:25 20VC: Why Now is the Time for the Application Layer | Why OpenAI & Anthropic Won't Win the App Layer | Why Startups Should be TokenMaxxing | Why VCs Should Reduce Weighting on Price & Ownership in an Age of AI with Mike Mignano, USV

Also on the record

Anastasios Angelopoulos · Aug 3, 2026

The claim that China is ahead because researchers work harder and enjoy policy, regulation and subsidy tailwinds overstates the difference; China has headwinds too, notably severe hardware constraints

The US has the best chip ecosystem in the world while China is hardware constrained, to the point of black-market chip imports

15:12 Chinas hardware headwinds offset its culture and policy tailwinds

Harry Stebbings · Aug 8, 2026

China's open-source models are a serious competitive threat that should alarm the US

The power and strength of Chinese open models

67:26 Us is behind and should be alarmed

Harry Stebbings · Aug 10, 2026

The gap between Chinese and US open-source models will be bigger in twelve months, because Chinese open-source providers are structurally advantaged

China anoints national champions like DeepSeek and Moonshot, concentrating state funding and sweeping aside regulation, while US open-source efforts must raise billions with a questionable business model while competing against OpenAI and Anthropic

32:47 Chinese state backing widens the gap

Your assistant can query this graph directly — 8 positions here, 19,646 across the corpus. Add 996.fm over MCP.