Is Claude 3.5 Sonnet a distilled or quantized version of a larger model?
11
Ṁ100Ṁ522resolved Feb 2
Resolved
NO1H
6H
1D
1W
1M
ALL
Resolves yes if Claude 3.5 Sonnet is distilled, quantized, or was made smaller via some other method, such that what 3.5 was distilled/quantized from is a larger (parameter number or similar) model.
Resolves no otherwise.
Must have good evidence one way or the other to resolve.
This question is managed and resolved by Manifold.
Market context
Get 
1,000 to start trading!
1,000🏅 Top traders
| # | Trader | Total profit |
|---|---|---|
| 1 | Ṁ56 | |
| 2 | Ṁ24 | |
| 3 | Ṁ9 | |
| 4 | Ṁ6 | |
| 5 | Ṁ6 |
Sort by:
https://darioamodei.com/on-deepseek-and-export-controls
> 3.5 Sonnet was not trained in any way that involved a larger or more expensive model
People are also trading
Related questions
Claude Sonnet 5 released this week?
1% chance
Which Claude Sonnet?
Claude Sonnet 4.6 METR 50% time horizon
Why is Claude 3.5 Sonnet such a good model for its size?
Have Claude Sonnet 4.5 model weights leaked before October 2025?
1% chance
Claude Sonnet 5 METR 50% time horizon
Will Manifold consider the next Claude Sonnet to be a frontier model?
81% chance
Will Claude Sonnet 5 exceed 85% on SWE-bench verified?
16% chance
Will the next Claude Sonnet be better than Claude 4.5 Opus at software engineering?
79% chance
How many (total) parameters does the Claude3 Haiku (the smallest variant) have? (resolve to log_10(N))