V* (V*)

A vision-centric benchmark for high-level multimodal reasoning and perception quality.

Top Models on V* — March 2026

As of March 2026, GLM-5V-Turbo leads the V* leaderboard with 89.0% , followed by Kimi K2.5 (84.3%) and Claude Opus 4.6 (66.5%).

3 modelsMultimodal & GroundedCurrentDisplay onlyUpdated April 1, 2026

According to BenchLM.ai, GLM-5V-Turbo leads the V* benchmark with a score of 89.0%, followed by Kimi K2.5 (84.3%) and Claude Opus 4.6 (66.5%). There is significant spread across the leaderboard, making this benchmark effective at differentiating model capabilities.

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

About V*

Year

2026

Tasks

Frontier multimodal reasoning tasks

Format

Vision-centric reasoning benchmark

Difficulty

Frontier multimodal

BenchLM tracks V* as a display-only frontier multimodal benchmark reference outside the current weighted schema.

GLM-5V-Turbo

BenchLM freshness & provenance

Version

V* 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.

Leaderboard (3 models)

#1GLM-5V-Turbo
89.0%
#2Kimi K2.5
84.3%
#3Claude Opus 4.6
66.5%

FAQ

What does V* measure?

A vision-centric benchmark for high-level multimodal reasoning and perception quality.

Which model scores highest on V*?

GLM-5V-Turbo by Zhipu AI currently leads with a score of 89.0% on V*.

How many models are evaluated on V*?

3 AI models have been evaluated on V* on BenchLM.

Last updated: April 1, 2026 · BenchLM version V* 2026

Weekly LLM Benchmark Digest

Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.

Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.