Full archive

Benchmark: Latest AI News

Search reporting on AI markets, models, products, companies, policy, and culture.

Showing 1-3 of 3Results matching "DeepSeek-R1"

Last week

2 stories

Aug 25, 2026

OpenAI’s Jalapeño Claims Up to 1.9x Efficiency Gain, but Leaves Nvidia Training Intact

The Broadcom co-developed inference processor is headed for deployment by year-end. Its published gains are meaningful, but they exclude model training and Nvidia’s Vera Rubin platform. OpenAI says Jalapeño, its custom chip co-developed with Broadcom, delivered up to 1.9 times higher throughput per kilowatt and 1.7 to 3.6 times lower end-to-end latency than Nvidia GB200 and GB300 systems in published tests.

3 min readRead story

Benchmark Infrastructure

Aug 25, 2026

OpenAI’s Jalapeño Claims 1.5–1.9x More AI Work Per Watt, Faces 2027 Scale Test

The custom chip gives OpenAI an early efficiency claim against available Nvidia systems, but production qualification, broader workloads and a small initial rollout will determine whether it becomes a meaningful serving platform. The company presented Jalapeño’s first benchmark results at Hot Chips, framing the chip as the opening generation of a longer-term silicon platform. Source: openai.com.

4 min readRead story

August 2026

1 story

Benchmark Models

Aug 21, 2026

Open Models Are Catching the Frontier Faster. Benchmark Scores Aren’t the Whole Contest.

SemiAnalysis sees the lag from closed-model breakthroughs to open-model benchmark parity shrinking across three AI eras. But a comparable model score still leaves a harder product question: who can turn capability into dependable agent work? In the reasoning and agentic eras, open AI models reached closed-model benchmark parity in months, according to SemiAnalysis ’s cross-era analysis.

4 min readRead story