Unified MemoryBenchmarks / Which Small AI Model Should You Run on a Mac? We Ran 312 Tests / Read images
vision-029
medium calendarcounting
The prompt
1 image attached (not shown).
user
How many events are on the calendar for Wednesday? Reply with only the number (no units or currency symbols).
What we expected
{
"answer": 3,
"accept": [],
"type": "number"
}What each model answered
LFM2.5-VL 3B
✗ fail · score 0 · 2.90 s
2
Scoring detail
{
"type": "number",
"parsed": 2
}MiMo-V2.6 Distill 9B
✗ fail · score 0 · 3.64 s
2
Scoring detail
{
"type": "number",
"parsed": 2
}MiniCPM-V 4.7 1B
✗ fail · score 0 · 1.78 s
2
Scoring detail
{
"type": "number",
"parsed": 2
}