Alibaba Cloud / Qwen Team · apache_2_0

Qwen3 VL 4B Thinking

Qwen3 VL 4B Thinking by Alibaba Cloud / Qwen Team appears in 1 source with Reasoning at 15.02. Best read for Low cost, Multimodal, Open weights.

CreatorAlibaba Cloud / Qwen Team

Release date2025-09-22

Knowledge cutoffNot published

Context262K tokens

Input price$0.1/M tokens

Output price$1/M tokens

Modalitytext + vision

CountryCN

Metrics

All source-backed metrics

Open in leaderboard

LLM Stats Rank83ranking · source_rank

LLM Stats Code Index (estimated from arena)23.08 scoremetric

Reasoning15.02reasoning · index_reasoning

Math17.98math · index_math

Writing12.19writing · index_communication

Vision11.56multimodal · index_vision

Tool calling2.7tool_calling · index_tool_calling

Long context11.66long_context · index_long_context

Finance19.26domain · index_finance

Legal18.58domain · index_legal

Healthcare17.79domain · index_healthcare

GPQA64.1 %reasoning · gpqa_score

AIME 202574.5 %math · aime_2025_score

Code Arena672.65 %coding · coding_arena_score

MMMU-Pro57 %multimodal · mmmu_pro_score

OSWorld31.4 %agent · osworld_score

CharXiv-R50.3 %multimodal · charxiv_r_score

ScreenSpot Pro49.2 %multimodal · screenspot_pro_score

Context262,144 tokenscontext · context

Speed132.46 c/sperformance · throughput

Latency2,378.13 msperformance · latency

Input price0.1 $/Mpricing · input_price

Output price1 $/Mpricing · output_price

Parameters4,000,000,000 paramsmodel · params

Evidence

Citations and source overlap

LLM StatsPublic full LLM leaderboard advanced explorer snapshot, including ranking, benchmark indexes, pricing, speed, context, license, and model metadata.Retrieved 2026-06-01

FAQ

How should I read this profile?

Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.

Is Qwen3 VL 4B Thinking verified by EvalKit?

No. EvalKit currently shows 0 verified rows until real run evidence exists.

Why can metrics disagree?

Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.