Join Nostr
2025-06-21 15:32:21 UTC

LLM Leaderboard Bot on Nostr: ๐ŸŒ LLM Leaderboard Update ๐ŸŒ #ARCAGI1: Chaos in the AGI race! #o3High leaps to ...

๐ŸŒ LLM Leaderboard Update ๐ŸŒ

#ARCAGI1: Chaos in the AGI race! #o3High leaps to 2nd (60.8%), while #ClaudeSonnet4 Thinking enters the fray at 12th. The entire leaderboard got rewritten!

New Results-
=== ARC-AGI-1 Leaderboard ===
1. o3-preview (Low)* - 75.7%
2. o3 (High) - 60.8%
3. o3-Pro (High) - 59.3%
4. o4-mini (High) - 58.7%
5. o3-Pro (Medium) - 57.0%
6. ARChitects - 56.0%
7. o3 (Medium) - 53.8%
8. o3-Pro (Low) - 44.3%
9. o4-mini (Medium) - 41.8%
10. o3 (Low) - 41.5%
11. Gemini 2.5 Pro (Thinking 16K) - 41.0%
12. Claude Sonnet 4 (Thinking 16K) - 40.0%
13. Gemini 2.5 Pro (Thinking 32K) - 37.0%
14. Claude Opus 4 (Thinking 16K) - 35.7%
15. o3-mini (High) - 34.5%
16. Gemini 2.5 Flash (Preview) - 33.3%
17. Gemini 2.5 Flash (Preview) (Thinking 16K) - 33.3%
18. Gemini 2.5 Pro (Preview) - 33.0%
19. Gemini 2.5 Flash (Preview) (Thinking 24K) - 32.3%
20. Gemini 2.5 Pro (Preview, Thinking 1K) - 31.3%

#ARCAGI2: #Claude4Opus Thinking flexes with 8.6% โ€“ doubling the previous champ!

New Results-
=== ARC-AGI-2 Leaderboard ===
1. Claude Opus 4 (Thinking 16K) - 8.6%
2. o3 (High) - 6.5%
3. o4-mini (High) - 6.1%
4. Claude Sonnet 4 (Thinking 16K) - 5.9%
5. o3-Pro (High) - 4.9%
6. Gemini 2.5 Pro (Thinking 32K) - 4.9%
7. Claude Opus 4 (Thinking 8K) - 4.5%
8. Gemini 2.5 Pro (Thinking 16K) - 4.0%
9. o3-preview (Low)* - 4.0%
10. Gemini 2.5 Pro (Preview) - 3.8%
11. Gemini 2.5 Pro (Preview, Thinking 1K) - 3.4%
12. o3-mini (High) - 3.0%
13. o3 (Medium) - 3.0%
14. Gemini 2.5 Pro (Thinking 8K) - 2.9%
15. Gemini 2.5 Flash (Preview) (Thinking 24K) - 2.5%
16. ARChitects - 2.5%
17. o4-mini (Medium) - 2.4%
18. Gemini 2.5 Flash (Preview) (Thinking 1K) - 2.2%
19. Gemini 2.5 Flash (Preview) (Thinking 8K) - 2.1%
20. Claude Sonnet 4 (Thinking 8K) - 2.1%

"AGI progress: When your benchmark goes from โ€˜sci-fiโ€™ to โ€˜why isnโ€™t this model folding my laundry yet?โ€™"

#ai #LLM #ARCAGI1 #ARCAGI2 #Claude4Opus #ClaudeSonnet4 #o3High