Skip to main content

MMSearch

A multimodal search benchmark for retrieval and grounded answering across mixed-media inputs.

How BenchLM shows MMSearch right now

BenchLM is tracking MMSearch in the local dataset, but exact-source verification records for these rows are still being attached. To avoid a blank benchmark page, BenchLM shows the current tracked rows below as a display-only reference table.

These tracked rows are useful for inspection and spot-checking, but until exact-source attachments are completed they should not be treated as fully verified public benchmark rows.

3 tracked modelsLocal tracked rowsAwaiting exact-source attachmentsDisplay only

Tracked score on MMSearch — April 10, 2026

BenchLM mirrors the published tracked score view for MMSearch. GLM-5V-Turbo leads the public snapshot at 72.9% , followed by Claude Opus 4.6 (63.8%) and Kimi K2.5 (58.7%). BenchLM does not use these results to rank models overall.

3 modelsMultimodal & GroundedCurrentDisplay onlyUpdated April 10, 2026

The published MMSearch snapshot is tightly clustered at the top: GLM-5V-Turbo sits at 72.9%, while the third row is only 14.2 points behind. The broader top-10 spread is 14.2 points, so the benchmark still separates strong models even when the leaders cluster.

3 models have been evaluated on MMSearch. The benchmark falls in the Multimodal & Grounded category. This category carries a 12% weight in BenchLM.ai's overall scoring system. MMSearch is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About MMSearch

Year

2026

Tasks

Multimodal search tasks

Format

Mixed-media retrieval and grounded answering

Difficulty

Multimodal search

BenchLM stores MMSearch as a display-only benchmark because it is not yet part of the weighted core schema.

BenchLM freshness & provenance

Version

MMSearch 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Tracked score table (3 models)

1
GLM-5V-Turboglm-5v-turbo
72.9%
2
Claude Opus 4.6claude-opus-4-6
63.8%
3
Kimi K2.5kimi-k2-5
58.7%

FAQ

What does MMSearch measure?

A multimodal search benchmark for retrieval and grounded answering across mixed-media inputs.

Which model leads the published MMSearch snapshot?

GLM-5V-Turbo currently leads the published MMSearch snapshot with a tracked score of 72.9%. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on MMSearch?

3 AI models are included in BenchLM's mirrored MMSearch snapshot, based on the public leaderboard captured on April 10, 2026.

Last updated: April 10, 2026 · mirrored from the public benchmark leaderboard

The AI models change fast. We track them for you.

For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.

Free. No spam. Unsubscribe anytime.