精选70Hacker News Best（web_list）模型发布/更新
DeepSeek V4 Flash 0731 发布
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731
DeepSeek·Jul 31, 2026·3 reasoning variants
Paper ↗Model ↗
At max effort, DeepSeek V4 Flash 0731 scores 89.0% on ARC-AGI-1 Semi-Private at $0.02 per task and 61.4% on ARC-AGI-2 Semi-Private at $0.04 per task.
ARC-AGI 2 leaderboard
DeepSeek V4 Flash 0731
ARC-AGI-1ARC-AGI-2
Verified scores
| Variant | ARC-AGI-1 | ARC-AGI-2 | ARC-AGI-3 |
|---|---|---|---|
| Max | 89.0% | 61.4% | — |
| High | 87.0% | 56.0% | — |
| Low | 84.0% | 46.0% | — |
Tasks & environments
Pass/fail per reasoning level across each benchmark.
ARC-AGI-2 Public Eval
120 tasks
| Task | Max✓✗ | High✓✗ | Low✓✗ |
|---|---|---|---|
| 0934a4d8 | ✗ | ✓ | ✗ |
| 135a2760 | ✓ | ✓ | ✓ |
| 136b0064 | ✓ | ✓ | ✓ |
| 13e47133 | ✗ | ✗ | ✗ |
| 142ca369 | ✗ | ✓ | ✗ |
| 16b78196 | ✗ | ✗ | ✗ |
| 16de56c4 | ✓ | ✓ | ✓ |
| 1818057f | ✓ | ✓ | ✓ |
| 195c6913 | ✓ | ✗ | ✗ |
| 1ae2feb7 | ✓ | ✓ | ✓ |
| 20270e3b | ✓ | ✓ | ✗ |
| 20a9e565 | ✗ | ✗ | ✗ |
| 21897d95 | ✗ | ✗ | ✗ |
| 221dfab4 | ✓ | ✓ | ✗ |
| 247ef758 | ✓ | ✓ | ✓ |
| 269e22fb | ✗ | ✗ | ✗ |
| 271d71e2 | ✗ | ✓ | ✗ |
| 28a6681f | ✓ | ✗ | ✗ |
| 291dc1e1 | ✓ | ✓ | ✗ |
| 2b83f449 | ✗ | ✗ | ✗ |
| 2ba387bc | ✓ | ✓ | ✓ |
| 2c181942 | ✓ | ✓ | ✗ |
| 2d0172a1 | ✗ | ✗ | ✗ |
| 31f7f899 | ✓ | ✓ | ✓ |
| 332f06d7 | ✓ | ✗ | ✗ |
| 35ab12c3 | ✓ | ✓ | ✓ |
| 36a08778 | ✓ | ✓ | ✓ |
| 38007db0 | ✓ | ✓ | ✓ |
| 3a25b0d8 | ✗ | ✗ | ✗ |
| 3dc255db | ✗ | ✗ | ✗ |
| 3e6067c3 | ✓ | ✗ | ✗ |
| 409aa875 | ✓ | ✓ | ✗ |
| 446ef5d2 | ✗ | ✗ | ✗ |
| 45a5af55 | ✓ | ✓ | ✓ |
| 4a21e3da | ✓ | ✗ | ✓ |
| 4c3d4a41 | ✗ | ✓ | ✓ |
| 4c416de3 | ✗ | ✗ | ✗ |
| 4c7dc4dd | ✗ | ✗ | ✗ |
| 4e34c42c | ✗ | ✗ | ✗ |
| 53fb4810 | ✓ | ✓ | ✓ |
| 5545f144 | ✗ | ✗ | ✗ |
| 581f7754 | ✗ | ✗ | ✗ |
| 58490d8a | ✓ | ✓ | ✓ |
| 58f5dbd5 | ✓ | ✓ | ✓ |
| 5961cc34 | ✓ | ✓ | ✓ |
| 5dbc8537 | ✓ | ✗ | ✗ |
| 62593bfd | ✗ | ✗ | ✗ |
| 64efde09 | ✗ | ✓ | ✗ |
| 65b59efc | ✓ | ✓ | ✓ |
| 67e490f4 | ✓ | ✓ | ✓ |
| 6e453dd6 | ✓ | ✓ | ✓ |
| 6e4f6532 | ✓ | ✗ | ✗ |
| 6ffbe589 | ✗ | ✗ | ✗ |
| 71e489b6 | ✓ | ✓ | ✗ |
| 7491f3cf | ✓ | ✗ | ✓ |
| 7666fa5d | ✓ | ✓ | ✗ |
| 78332cb0 | ✗ | ✗ | ✗ |
| 7b0280bc | ✓ | ✗ | ✗ |
| 7b3084d4 | ✓ | ✓ | ✗ |
| 7b5033c1 | ✗ | ✗ | ✓ |
| 7b80bb43 | ✗ | ✗ | ✗ |
| 7c66cb00 | ✗ | ✗ | ✓ |
| 7ed72f31 | ✓ | ✓ | ✗ |
| 800d221b | ✗ | ✗ | ✗ |
| 80a900e0 | ✓ | ✓ | ✓ |
| 8698868d | ✗ | ✗ | ✗ |
| 88bcf3b4 | ✗ | ✗ | ✗ |
| 88e364bc | ✗ | ✗ | ✗ |
| 89565ca0 | ✓ | ✗ | ✓ |
| 898e7135 | ✓ | ✓ | ✓ |
| 8b7bacbf | ✗ | ✗ | ✗ |
| 8b9c3697 | ✓ | ✗ | ✗ |
| 8e5c0c38 | ✓ | ✗ | ✓ |
| 8f215267 | ✓ | ✓ | ✓ |
| 8f3a5a89 | ✓ | ✓ | ✓ |
| 9385bd28 | ✓ | ✓ | ✗ |
| 97d7923e | ✓ | ✗ | ✓ |
| 981571dc | ✓ | ✓ | ✓ |
| 9aaea919 | ✓ | ✓ | ✓ |
| 9bbf930d | ✗ | ✗ | ✗ |
| a251c730 | ✓ | ✓ | ✓ |
| a25697e4 | ✗ | ✗ | ✗ |
| a32d8b75 | ✗ | ✗ | ✗ |
| a395ee82 | ✓ | ✓ | ✓ |
| a47bf94d | ✗ | ✓ | ✓ |
| a6f40cea | ✓ | ✓ | ✓ |
| aa4ec2a5 | ✓ | ✓ | ✓ |
| abc82100 | ✗ | ✗ | ✗ |
| b0039139 | ✓ | ✗ | ✓ |
| b10624e5 | ✓ | ✓ | ✓ |
| b5ca7ac4 | ✓ | ✓ | ✓ |
| b6f77b65 | ✓ | ✓ | ✗ |
| b99e7126 | ✓ | ✓ | ✓ |
| b9e38dc0 | ✗ | ✗ | ✗ |
| bf45cf4b | ✓ | ✓ | ✓ |
| c4d067a0 | ✓ | ✓ | ✓ |
| c7f57c3e | ✓ | ✗ | ✗ |
| cb2d8a2c | ✓ | ✓ | ✗ |
| cbebaa4b | ✓ | ✓ | ✗ |
| d35bdbdc | ✗ | ✗ | ✗ |
| d59b0160 | ✓ | ✓ | ✓ |
| d8e07eb2 | ✓ | ✓ | ✓ |
| da515329 | ✗ | ✗ | ✗ |
| db0c5428 | ✓ | ✓ | ✓ |
| db695cfb | ✓ | ✓ | ✓ |
| dbff022c | ✓ | ✓ | ✗ |
| dd6b8c4b | ✗ | ✗ | ✗ |
| de809cff | ✗ | ✗ | ✗ |
| dfadab01 | ✗ | ✓ | ✗ |
| e12f9a14 | ✗ | ✗ | ✗ |
| e3721c99 | ✗ | ✗ | ✗ |
| e376de54 | ✓ | ✓ | ✗ |
| e8686506 | ✓ | ✓ | ✓ |
| e87109e9 | ✓ | ✓ | ✗ |
| edb79dae | ✗ | ✓ | ✓ |
| eee78d87 | ✓ | ✗ | ✗ |
| f560132c | ✗ | ✗ | ✗ |
| f931b4a8 | ✓ | ✓ | ✓ |
| faa9f03d | ✗ | ✗ | ✗ |
| fc7cae8d | ✗ | ✓ | ✗ |
ARC-AGI-1 Public Eval
400 tasks
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力