# SWE-bench Multimodal scores barely drop without images, study finds

> Original title: this post is such a banger. re: SWE-bench Multimodal

- Source: Hamel Husain
- Published: 2026-11-03T18:39:11.000Z
- Bad Influence: https://badinfluence.news/items/mosqaelsfnwe24gaw52du7bv4
- Original: https://x.com/HamelHusain/status/2117688053129171206

## Summary

Hamel Husain praises a post by @dejavucoder arguing that SWE-bench Multimodal is not truly multimodal. Removing the images barely changes scores, and most tasks can be solved from the issue text and code alone.

## Post

this post is such a banger.  re: SWE-bench Multimodal

"Removing the images barely changes the score ... Most tasks can be solved from the issue text and the code alone"

## Quoted post @dejavucoder

> curious people and people interested in llm evals, benchmarks, high data circles, please wake up. my blog post "SWE-bench Multimodal v2 is not multimodal" is up now. 
> 
> https://sankalp.bearblog.dev/swe_multimodal/

https://x.com/dejavucoder/status/2112910513982292256