Compare 500+ AI models across vision, tool use, function calling, JSON mode, structured output, reasoning, streaming, and context windows. Claude Opus 5 (Adaptive Reasoning, Max Effort, Intelligence Index 60.7, $5/$25), Claude Opus 5 (Xhigh Effort, II 60.1), and Claude Fable 5 (II 59.9, $10/$50) lead today — check the matrix below for vision, 1M+ context, JSON mode, and reasoning support across every frontier model. Built for developers picking a model for chat, code, agents, image understanding, or long-document tasks. For hands-on rankings, see our live LLM benchmark leaderboard, AI model speed rankings, best coding plan subscriptions, GPQA/AIME/HLE reasoning comparison, and PinchBench real-world agent scores.
Compare 0 AI models across 14 features — vision, tool use, context windows, JSON mode, reasoning, and more. Find exactly which capabilities each model supports.
| Model | Context | Max Out | $/M In | StreamingStreaming | Vision / Image InputVision | Audio InputAudio | Video InputVideo | File UploadFile | Image GenerationImage | Tool Use / Function CallingTool | JSON / Structured OutputJSON | Reasoning / CoTReasoning | Web SearchWeb | Log ProbabilitiesLog | Reproducible (Seed)Reproducible |
|---|
Most flagship models from OpenAI (GPT-5.4), Anthropic (Claude 4.6), and Google (Gemini 3) support image input. Check the Vision column in the matrix above for the full list.
Function calling is widely supported by OpenAI, Anthropic, Google, and Mistral models. Look for the Tool Use column. Models with this feature can call external APIs and tools you define.
As of 2026, some models offer context windows of 1M+ tokens (Google Gemini, some OpenAI models). Larger context windows let you process entire codebases or long documents in a single request.
Most modern models support JSON mode or structured outputs. This ensures the model returns valid JSON matching your schema — essential for API integrations and data extraction tasks.
Nearly all major API models support streaming, which sends tokens as they are generated. This provides a much better user experience for chat applications with faster perceived response times.