What will be true of DeepSeek's r2 model?
23
Ṁ1915Apr 1
92%
Some version of it will score better on GPQA than o1-preview (73% pass@1)
74%
Its base model will be DeepSeek-V3
63%
It will support a longer context length than DeepSeek-r1
55%
DeepSeek will report it scores >= 65% on SWE-Bench Verified
54%
It will score >= 78% on LiveBench when first added (R1 = 71.5% on March 12)
45%
Manifold thinks it is better than expected in a poll asking if it is better or worse than expected
37%
It will have a variant with more parameters than DeepSeek-r1 (not necessarily active)
12%
Reaches Highest Arena Score on Chatbot Arena
everything N/As if a model called R2 created by DeepSeek is not released in 2025
This question is managed and resolved by Manifold.
Get
1,000and
3.00
1,0003.00Related questions
Related questions
Will DeepSeek's next reasoning model be called R3?
3% chance
Will ANY DeepSeek model cause or materially enable a catastrophic risk by 2027?
7% chance
When will DeepSeek release V4?
Will DeepSeek R2 be open source?
83% chance
When will DeepSeek release R2?
"Holy shit!" -> my reaction to deepseek r1. Will I feel the same about any AI developments in the next 5 months?
35% chance
When will Deepseek R2 be released?
-
How much did DeepSeek-V3 cost to train?
Will DeepSeek's next reasoning model be open-sourced?
79% chance
Will I believe in 1 year that DeepSeek R1 was substantially trained via distillation of a US model?
64% chance