Skip to content

Compare models

Put up to 4 models beside each other — token prices, context windows, capabilities and provider, from the same catalogue the model pages read.

  1. Nemotron 3.5 Super VL (Free)NVIDIARemove
  2. Gemini 3.5 TranscribeGoogleRemove
  3. Qwen3.8 27BAlibabaRemove
nemotron-3.5-super-vl:free vs gemini-3.5-transcribe vs qwen3.8-27b
AttributeNemotron 3.5 Super VL (Free)nemotron-3.5-super-vl:freeGemini 3.5 Transcribegemini-3.5-transcribeQwen3.8 27Bqwen3.8-27b
Pricing
Input$0 / 1M— Not priced per input token$0.45 / 1M
Output$0 / 1M— Not priced per output token$3.20 / 1M
Cache Write$0 / 1M——
Cache Read$0 / 1MNot applicable$0.45 / 1M
Cache Write (5m)—Not applicable$0.45 / 1M
Cache Write (1h)—Not applicable$0.45 / 1M
Web Search—$0 / 1M$0 / 1M
Context
Max context1M98.3K262K
Max outputN/AN/AN/A
Capabilities
VisionYesYesYes
Function CallingYesYesYes
JSON ModeYesNoYes
StreamingYesNoYes
Catalogue
ProviderNVIDIAGoogleAlibaba
Categorychatvoicechat
Charge typeFreePay As You GoPay As You Go
Released—2026-09-25—
Description
SummaryNVIDIA Nemotron 3.5 Super VL is an open multimodal Mixture-of-Experts (MoE) model featuring 120B total parameters with 12B activated per token. Built on a hybrid Mamba–Transformer architecture with multi-token prediction, it supports text, images, documents, GUI screenshots, and video inputs, along with a 1M-token context window for complex, long-horizon workflows. Optimized for multimodal reasoning, visual understanding, and agentic tool use, the model supports configurable reasoning modes and token budgets. It is well suited for document intelligence, web research, coding validation, customer service, GUI analysis, and video understanding, enabling efficient, long-running multimodal agent applications.Google Gemini 3.5 Transcribe speech-to-text. Billed per input and output token.Qwen3.8 27B is an open-weight dense vision-language model from Qwen, designed for coding, professional knowledge work, research, and multimodal interaction. It combines strong text and visual understanding with capabilities optimized for sustained, real-world agentic tasks. The model supports flexible thinking modes that can be enabled for deeper reasoning or disabled for faster execution, making it well suited for long-running agents, multimodal workflows, coding assistants, and cost-conscious self-hosted deployments.