Qwen3 VL 8B Thinking
by Qwentext+image->text1 endpoint
Neura Intelligence Index
49.8/ 100
Rank
#152—
Confidence
mediumBenchmarks
7(18% coverage)
Overview
Qwen3-VL-8B-Thinking is the reasoning-optimized variant of the Qwen3-VL-8B multimodal model, designed for advanced visual and textual reasoning across complex scenes, documents, and temporal sequences. It integrates enhanced multimodal alignment and...
Capabilities
Text generation
Image understanding
Tool use / Function calling
Structured output
Extended reasoning
Long context
Modalities
Input
ImageText
Output
Text
Technical Specifications
Context Window
256K tokens
Max Output
32.8K tokens
Tokenizer
Qwen3
Uptime (24h)
10000.0%
Parameters
9000000000B
License
apache_2_0Open Weights
Supported Parameters
include_reasoninglogprobsmax_tokenspresence_penaltyreasoningresponse_formatseedstructured_outputstemperaturetool_choicetoolstop_logprobstop_p
Benchmark Performance
Performance Over Time
Pricing
Per 1M tokens
Input$0.12
Output$1.36