DeepSeek逼近编程用量榜首，白宫审议模型评测框架
【DeepSeek逼近编程用量榜首，白宫审议模型评测框架】金色财经报道，8月4日，DeepSeek V4 Flash在Kilo Code近24小时榜升至第二，与榜首Step 3.7 Flash只差10亿词元；OpenRouter周榜前五名继续全部来自中国团队。与此同时，白宫将与AI企业审议前沿模型发布前的自愿评测框架，真实调用量与模型发布规则成为今天的两条主线。(金十)()
| 排名 | # | 交易平台 | 24H成交额 | 成交涨跌幅 | 活跃度 | 交易平台 | 活跃度 |
|---|---|---|---|---|---|---|---|
| 1 | $185.59亿 | -11.59% | 86.7% | 86.7% | |||
| 2 | $26.37亿 | 0.44% | 68.6% | 68.6% | |||
| 3 | $404.63亿 | -8.05% | 85.4% | 85.4% | |||
| 4 | $9.01亿 | -15% | 69.3% | 69.3% | |||
| 5 | $109.78亿 | -6.6% | 65.3% | 65.3% | |||
| 6 | $12.64亿 | -8.54% | 53.3% | 53.3% | |||
| 7 | $2.06亿 | 19.4% | 65.3% | 65.3% | |||
| 8 | $7.58亿 | -8.93% | 68.4% | 68.4% | |||
| 9 | $2.87亿 | -21.41% | 56.9% | 56.9% | |||
| 10 | $72.07亿 | 0.51% | 71.1% | 71.1% | |||
| 11 | $4.36亿 | 5.31% | 36.2% | 36.2% | |||
| 12 | $3,097.39万 | -18.67% | 59.5% | 59.5% | |||
| 13 | $216.37亿 | -0.43% | 31% | 31% | |||
| 14 | $112.87亿 | 1.26% | 76.2% | 76.2% | |||
| 15 | $1.44亿 | -9.44% | 66.9% | 66.9% | |||
| 16 | $107.10亿 | -12.51% | 55.2% | 55.2% | |||
| 17 | $11.94亿 | -6.43% | 26% | 26% | |||
| 18 | $16.48亿 | 26.65% | 46.9% | 46.9% | |||
| 19 | $1.07亿 | -10.3% | 31.5% | 31.5% | |||
| 20 | $1,688.77万 | -20.7% | 58.2% | 58.2% |
