Unverified50% confidenceBenchmarkExact time

On a TPU v4-32 pod, nanoAlphaZero can train a chess model with an Elo rating above 2700 within 24 hours.

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

8/6/2026

First Seen

Valid until: 11/4/2026

Sources

nanoAlphaZero：单文件JAX实现24小时训练2700分棋力

Related Claims

UnverifiedTPU v4首次部署了光交换互联（OCS）技术，使数千颗芯片能够组成超大规模Pod进行分布式训练58% similar UnverifiedExoLabs can already run the 550-billion-parameter Nemotron 3.0 on 4 Spark units at 30 tokens per second.58% similar Unverified去噪塔仅用2.1万亿token训练，而原始Nemotron Nano使用了25万亿token，因此该模型是解码方式的转换而非更聪明的模型57% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/699042

API

curl https://kongchang.com/api/v1/knowledge/claims/699042

MCP

get_claim(id=699042)