LOADING COMPARISON
GROUNDED IN REAL HUMAN CONVERSATIONSYOUR BENCH
PICK MODELS TO COMPARE
Now pick a second model to put against Llama 3.1 70B Instruct. Fair fights only, so these are all chat and coding models.
SELECTED: LLAMA 3.1 70B INSTRUCT
Grok 4.6MULTIMODAL#1 · 42,001 VOICES
7.7VIBE
Qwen3.8 27BMULTIMODAL#2 · 13,970 VOICES
7.6VIBE- DiscoveryTEXT
#3 · 9,856 VOICES
7.6VIBE 
GLM 5.3MULTIMODAL#4 · 9,813 VOICES
7.5VIBE- Step 2TEXT
#7 · 3,772 VOICES
7.5VIBE 
DeepSeek V4 FlashTEXT#9 · 6,334 VOICES
7.4VIBE- Step 1TEXT
#11 · 3,795 VOICES
7.4VIBE - Qwen3.8 FlashMULTIMODAL
#12 · 2,897 VOICES
7.4VIBE - Dots3 NoteMULTIMODAL
#13 · 2,634 VOICES
7.4VIBE - Apertus v1.5 70BTEXT
#14 · 2,526 VOICES
7.4VIBE - GLM 5.3 FlashMULTIMODAL
#16 · 4,443 VOICES
7.3VIBE - DeepSeek V4 Flash Vision ExpMULTIMODAL
#17 · 4,361 VOICES
7.3VIBE