返回
原文
Chubby♨️@kimmonismus
64AI 编辑部评分，满分 100
2026-07-31 17:38· 22分钟前
跳到正文
AI 摘要

DeepSeek v4 Flash 在 Artificial Analysis 智能指数上得分 50，较上代提升 10 分，单位任务价格目前无人能敌。其 DeepSWE 达 54.4%，超越 GLM-5.2 和 4 Pro，几乎持平 Opus 4.8；TerminalBench 82.7%，同样逼近 Opus 4.8。

The new DeepSeek v4 Flash scored 50 points on the Artificial Analysis Intelligence Index， a jump of 10 points compared to the previous version.

In terms of price per task， the new model is currently unbeatable.

Even if the performance doesn't quite meet expectations， the new Flash is still an incredibly affordable and very good model.

Nobody would have predicted two years ago how good and affordable these models would be.

Chubby♨️This is insane! not kidding at all. DeepSeek v4 *Flash* super close to Opus 4.8, Insane upgrade - DeepSWE 54.4%, outperforming GLM-5.2, its 4 Pro version by a l...
DeepSeek模型发布评测/基准
在 X 查看原推导出 Markdown
返回精选列表
64导出 Markdown
Chubby♨️ · @kimmonismus · X·2026-07-31 17:38·22分钟前
在 X 看原推· x.com
AI 摘要

DeepSeek v4 Flash 在 Artificial Analysis 智能指数上得分 50，较上代提升 10 分，单位任务价格目前无人能敌。其 DeepSWE 达 54.4%，超越 GLM-5.2 和 4 Pro，几乎持平 Opus 4.8；TerminalBench 82.7%，同样逼近 Opus 4.8。

The new DeepSeek v4 Flash scored 50 points on the Artificial Analysis Intelligence Index， a jump of 10 points compared to the previous version.

In terms of price per task， the new model is currently unbeatable.

Even if the performance doesn't quite meet expectations， the new Flash is still an incredibly affordable and very good model.

Nobody would have predicted two years ago how good and affordable these models would be.

Chubby♨️This is insane! not kidding at all. DeepSeek v4 *Flash* super close to Opus 4.8, Insane upgrade - DeepSWE 54.4%, outperforming GLM-5.2, its 4 Pro version by a l...
DeepSeek模型发布评测/基准
在 X 查看原推x.com