in vision, Claude Fable is on par with an *old 3B active Qwen* (Qwen-Flash is basically just hosted Qwen3.5-35B-A3B) that's all you get as a spillover from general scale
4:07 PM · Jun 9, 2026 · 4.3K Views
Claude Fable-5 scores 20.0 on Eyebench-V3 vision benchmark, tying Qwen3.5-Flash and barely beating Claude Opus 4.7
Kalomaze blamed the results on a frozen vision encoder
9923135.9K
Sentiment
Users are frustrated with Claude Fable-5 tying a competitor on the Eyebench-V3 vision benchmark, complaining that Anthropic is not trying hard enough on vision while preferring other models.
Pos
0.0%
Neg
100.0%
3 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS909BOOKMARKS2REPLIES2
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTex
yeah just one benchmark, I'm exaggerating but this is directionally true. They're not even trying
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTex
in vision, Claude Fable is on par with an *old 3B active Qwen* (Qwen-Flash is basically just hosted Qwen3.5-35B-A3B) that's all you get as a spillover from general scale
3hViews 909Likes 12Bookmarks 2
LIKES14
kalomaze@kalomaze
@teortaxesTex God is it a frozen vision encoder or something GOD why is Google mogging them so hard on this
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTex
in vision, Claude Fable is on par with an *old 3B active Qwen* (Qwen-Flash is basically just hosted Qwen3.5-35B-A3B) that's all you get as a spillover from general scale
3hViews 523Likes 14Bookmarks 0
kalomaze@kalomaze
@teortaxesTex i trust gemma4 26b more for vision than sonnets
kalomaze@kalomaze
@teortaxesTex God is it a frozen vision encoder or something GOD why is Google mogging them so hard on this
3hViews 293Likes 9Bookmarks 1
Offset Zero@offsetx0
@teortaxesTex It is better in spatial reasoning. Vision is not their main property as of now, but I think this benchmark is heavily concentrated on a very narrow AI blindspot, which is more of a vision encoder benchmark than the model.
3hViews 8