NVIDIA Swaps Harness, Lifts AI Agent Score from 30% to 100%
Custom harness on Claude Opus 5 hits perfect score on ARC-AGI-3 — the scaffolding, not the model, made the difference
160
NVIDIA
A company that makes the chips that power AI. Rather than a competitor in the chatbot race, it's the one selling picks and shovels to everyone in it. Its flagship products are GPUs and the AI infrastructure software built on top of them. In August 2026, it began officially supporting its cloud gaming service GeForce NOW on Linux, and around the same time, it published a method for isolating Kubernetes tenants on shared GPUs. In autonomous driving, it introduced a 34-billion-parameter reasoning model, and in robot policy learning, it presented a video-based learning approach. It also tried to expand the AI ecosystem for handling the physical world through its open world model 'Cosmos 3'.
Current rank (1M)
5
- Change
- Prev.
- 4
Rank over the last 30 days
2026.08.18 – 2026.09.16 · High 3 · Low 5 · Now 5
NVIDIA · Related stories
Starcloud raises additional $250M for orbital data centers
Valuation jumps to $2.3 billion... securing capital to weather SpaceX rocket supply crunch
00
Tencent's Zhuque Lab Open-Sources AI Agent/MCP Security Scanner
A red-team tool checks everything from agent, skill, and MCP server vulnerabilities to LLM jailbreaks on a single platform
10
Modular, Now Owned by Qualcomm, Fully Open-Sources Mojo Language
Compiler included under Apache 2.0 — MiniMax already serving billions of tokens per minute
60
Liquid AI's 300M Draft Model Speeds Up Decoding by Up to 3.18x
DSpark released for three LFM2.5 models — greedy output identical to the original, MacBook throughput up 61→139 tokens/sec
20
NVIDIA releases Switchyard, an LLM routing proxy for coding agents
A Rust proxy that connects Claude Code and Codex to open-source models as-is — not yet production-ready
00
NVIDIA's 300 Verified Skills Lift Correctness by 41 Points
NVIDIA published results measured on the same tasks and same models, changing only whether a skill was installed
00
Factory Commits $100M to Partner Network, Pushes to Scale Software Factories
Two new tracks—strategists and engineers—for consulting and SI partners, first cohort launching this quarter
00
MiniMax unveils music model that generates full 5-minute songs from lyrics alone
Open-weight MiniMax-Music3 generates a complete 32kHz stereo track in one pass from just two inputs: lyrics and a structured caption
30
NVIDIA invests $1.5 billion in SoftBank-linked data center developer
In exchange for exclusive chip supply to Ohio OpenAI data center, NVIDIA offers up to $105 billion in credit, not $10.5 billion
20
Higgsfield's First AI Feature Film Screened Alongside Pixar, Disney at SIGGRAPH
"KÖK BÖRÜ," based on a Kazakh folktale, was made with Seedance 2.0 and released with its full prompts
280
Groq raises $350 million, drops in-house chip plans to become an NVIDIA customer
Valuation halved from $6.9 billion last September to $3.5 billion, as the company fully pivots to a neocloud model leasing out NVIDIA GPUs
20
NVIDIA · Recent announcements
- 251엔비디아, 허깅페이스 129억달러에 인수
- 64410 million downloads for NVIDIA Warp
- 460Already running an inference engine? So where does NVIDIA Dynamo fit in?
- 170Can a security defense catch an attack it hasn’t seen before?
- 72Ask the Experts: NVIDIA NeMo Switchyard | Nemotron Labs
- 73Run Your Polars Code on Multiple GPUs | Live with cuDF Polars
- 9nvidia/DeepSeek-V4-Pro-0813-NVFP4 · text-generation
- 2nvidia/Muse-Glimmer-30B-NVFP4 · image-text-to-text
- –NVIDIA and CrowdStrike Strengthen Agentic Cybersecurity Frontier
- –GeForce NOW Gives Gamers More Ways to Play at Gamescom 2026
- –Delivering Vera: NVIDIA’s First CPU Built for Agents Is Shipping Now
- –NVIDIA NVLink Fusion Expands With NVHBM Custom High-Bandwidth Memory