Superpower Daily
University of Hawaiʻi Gets $2M to Test AI Early Warnings for Fo...Fervo Targets October Start for Potential First U.S. Commercial...InferenceX Ranks B300 First for Qwen3.5 at 15,815 Tokens per Se...Fix Victoria Uses AI Ads Ahead of Victoria’s Election While Its...Artificial Analysis Clarifies Astra’s 67-Point Coding Score Is...More Than $45 Million in Data-Center Ads Pushes AI Infrastructu...University of Hawaiʻi Gets $2M to Test AI Early Warnings for Fo...Fervo Targets October Start for Potential First U.S. Commercial...InferenceX Ranks B300 First for Qwen3.5 at 15,815 Tokens per Se...Fix Victoria Uses AI Ads Ahead of Victoria’s Election While Its...Artificial Analysis Clarifies Astra’s 67-Point Coding Score Is...More Than $45 Million in Data-Center Ads Pushes AI Infrastructu...
Swipe

Full archive

Benchmark: Latest AI News

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

Category

More filters
Benchmark
Showing 1-6 of 6Results matching "B300"

Today

2 stories
BenchmarkInfrastructure

InferenceX Ranks GB200 First for MiniMax M3, While MI355X Has Lower Token Cost

The live AgentX result gives operators a comparable speed-and-price snapshot at one responsiveness target, but the ranking can shift with serving software and does not settle the wider configuration trade-offs. Infrastructure teams serving MiniMax M3 now face a clean split in InferenceX ’s latest ranking: GB200 NVL72 delivers more throughput, while MI355X carries the lower reported price per million tokens.

3 min readRead story

Earlier this week

1 story

Last week

2 stories
BenchmarkInfrastructure

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
BenchmarkInfrastructure

NVIDIA Says Vera Rubin Delivers 30x More Agentic Throughput per Megawatt, Pending Review

The preliminary comparison shifts attention from fixed prompt benchmarks toward the power required to keep long, tool-using AI sessions responsive. NVIDIA says Vera Rubin NVL72 delivered up to 30 times the throughput per megawatt of GB300 NVL72 for an agentic AI workload at 160 output tokens per second per user. NVIDIA reports the DeepSeek V4 Pro 1.6T comparison on the AgentX workload.

3 min readRead story

August 2026

1 story