Claude Haiku 4.5
Fast and inexpensive model in the Claude family, ideal for high-volume use cases.
- Input price
- $1.00
- Output price
- $5.00
- Context window
- 200K
- Input
- Text, Image
Browse every tracked AI model with prices, specs, and sources.
Fast and inexpensive model in the Claude family, ideal for high-volume use cases.
Established top-tier Opus 4 model with 1M context and very strong reasoning.
High-performance Opus 4 family model for demanding reasoning and coding tasks with 1M context window.
Anthropic's current top model for complex agentic tasks, coding, and deep reasoning. Uses a new tokenizer.
Proven Sonnet model for scaling at moderate cost, with 1M context.
Balanced mid-tier model with 1M context, ideal for production workloads.
Cost-effective Cohere model for RAG applications with 128k context.
Cohere's high-performance model for enterprise RAG, tool use, and multilingual applications.
DeepSeek's current main model with 1M context window and very aggressive pricing.
Pro variant of DeepSeek V4 with higher capability. Currently 75% off until 2026-05-31.
Fast, balanced multimodal Gemini 2.5 model with 1M context window.
Most affordable Gemini 2.5 variant, optimized for high volume at low cost.
Established top-tier Gemini 2.5 model with tiered pricing based on prompt length.
Very affordable Gemini 3.1 variant with full multimodality.
Current fast multimodal model in the Gemini 3 family with native text, image, video, and audio support.
Balanced GPT-5 family model with strong price-performance for general applications.
Cost-efficient variant of GPT-5.4 for well-defined tasks at high speed.
OpenAI's current flagship model for complex reasoning, coding, and agentic tasks across all domains.
Premium variant of GPT-5.5 for the most demanding reasoning tasks with maximum accuracy.
Current Grok model from xAI with 1M context window and competitive pricing.
Larger Meta Llama 4 open-source model with MoE architecture, multimodal input and 1M context window.
Open-source Meta model with mixture-of-experts architecture (17B active parameters, 109B total) and exceptional context window up to 10M tokens.
Mistral AI's flagship model for reasoning, code, JSON and chat with 128k context, excellent across many languages.
Balanced mid-tier Mistral model offering reasoning and multimodal performance at much lower cost than enterprise models.
Compact Mistral open-source model (24B params) with vision support, tool use and excellent efficiency for coding and STEM.
Specialized open-source coding model with 1M context window and MoE architecture (480B parameters, 35B active).
Alibaba's largest Qwen3 model with 262k context window, strong reasoning and multilingual performance.
Preview of the top model in the Gemini 3.1 family with tiered pricing based on prompt length.