Models

Explore AI model pricing, capabilities, endpoints, and vendor coverage from one production catalog.

OpenAI

GPT-6.2 Astra

gpt-6.2-astra

GPT-6.2 Astra is OpenAI's flagship GPT-6.2 model, built for the hardest projects. It suits real-world engineering, professional work, and scientific research, with stronger visual reasoning for websites, slideshows, and 3D designs. OpenAI reports new eval highs at less than a third of GPT-6 Astra's cost per task.

Total Context

1.1M

Max Output

128K

Released

Nov 17, 2026

Input$8 / M tokens
Output$40 / M tokens
Cache Read$0.4 / M tokens

OpenAI

GPT-6.2 Astra Pro

gpt-6.2-astra-pro

GPT-6.2 Astra Pro uses the same underlying model as GPT-6.2 Astra, served with reasoning.mode set to pro for higher-quality responses on complex tasks. Cost note: pro mode uses far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6.2 Astra and takes much longer. Use it for hard, high-stakes problems where extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6.2 Astra instead.

Total Context

1.1M

Max Output

128K

Released

Nov 17, 2026

Input$8 / M tokens
Output$40 / M tokens
Cache Read$0.4 / M tokens

Tencent

Hy4 flash

hy4-flash

hy4-flash has roughly one-quarter the parameters of the flagship model. With lower pricing, it matches flagship open-source performance on everyday tasks — well suited to coding assistance, agent workflows, document and chart understanding, and high-volume daily office work.

Total Context

1M

Max Output

65.5K

Released

Nov 16, 2026

Input$0.035 / M tokens
Output$0.1405 / M tokens
Cache Read$0.0015 / M tokens

Moonshot

Kimi K3.1 Pro

kimi-k3.1-pro

Kimi K3.1 Pro is Moonshot AI's flagship open-weight multimodal reasoning model and a post-trained successor to Kimi K3, featuring a 1M-token context window. It is built for long-horizon coding, 3D creation, financial research, and end-to-end agentic knowledge work, with native image and video understanding.

Total Context

1M

Max Output

131.1K

Released

Oct 23, 2026

Input$2 / M tokens
Output$10 / M tokens
Cache Read$0.2 / M tokens

Alibaba

Qwen4-Max

qwen4-max

Qwen4-Max is Tongyi Qianwen's new-generation flagship multimodal model, combining text generation, image and video understanding, and deep reasoning. It is suited for coding, office work, front-end design, and complex agent tasks. The model supports web search, tool calling, and structured output, and is compatible with OpenAI and Anthropic APIs, helping developers build intelligent applications with multimodal understanding and task execution capabilities.

Total Context

1M

Max Output

262.1K

Released

Nov 16, 2026

Input$1.425 / M tokens
Output$4.275 / M tokens
Cache Read$0.0713 / M tokens

OpenAI

GPT-6.2 Sol

gpt-6.2-sol

GPT-6.2 Sol is an upgrade to OpenAI's GPT-6.1 Sol, positioned below the flagship GPT-6.2 Astra in the GPT-6.2 series. It is an efficient, versatile model for everyday work, suited to multistep agentic workflows and clear, natural writing across technical and non-technical tasks.

Total Context

1.1M

Max Output

128K

Released

Nov 17, 2026

Input$2 / M tokens
Output$10 / M tokens
Cache Read$0.2 / M tokens

OpenAI

GPT-6.2 Sol Pro

gpt-6.2-sol-pro

GPT-6.2 Sol Pro is the same underlying model as GPT-6.2 Sol, served with reasoning.mode set to pro for higher-quality responses on complex tasks. Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6.2 Sol and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6.2 Sol instead.

Total Context

1.1M

Max Output

128K

Released

Nov 17, 2026

Input$2 / M tokens
Output$10 / M tokens
Cache Read$0.1 / M tokens

TypeSafe

Jev 1.14

jev-1.14

Jev 1.14 is TypeSafe's System One model for fast, structured decision-making in software. It evaluates text and application state against developer-defined questions, returning choices, yes/no probabilities, and scores that applications can use directly. Built for classification, routing, and automated workflows, it helps developers turn natural-language judgments into actionable outputs and supports evaluating multiple questions in parallel within a single request.

Total Context

32K

Max Output

N/A

Released

Oct 14, 2026

Input$0.042 / M tokens
Output$0 / M tokens

TypeSafe

Jev Latest

jev-latest

Jev Latest provides access to the newest release of TypeSafe’s Jev model, a System One decision model built for fast, structured judgments in software. It evaluates text and application state against predefined questions, returning choices, yes/no probabilities, and scores that code can use directly. The Jev API supports classification, routing, and automated workflows, while Latest follows new releases in the Jev family.

Total Context

32K

Max Output

N/A

Released

Sep 18, 2026

Input$0.042 / M tokens
Output$0 / M tokens

OpenAI

GPT-6.1 Sol

gpt-6.1-sol

GPT-6.1 Sol is an advanced reasoning model built for complex problem solving, coding, and agent-based applications. It is designed to deliver strong reasoning performance while maintaining a practical balance between intelligence, speed, and cost. Unlike traditional chat models that focus mainly on generating responses, GPT-6.1 Sol is optimized for tasks that require deeper thinking, including multi-step reasoning, code generation, debugging, data analysis, and following complex instructions. GPT-6.1 Sol supports reliable reasoning workflows, structured outputs, and tool-based applications, making it suitable for developers building AI assistants, coding agents, automation systems, and production AI applications.

Total Context

1.1M

Max Output

128K

Released

Sep 29, 2026

Input$2 / M tokens
Output$10 / M tokens
Cache Read$0.1 / M tokens

Anthropic

Claude Haiku 5.5

claude-haiku-5.5

Claude Haiku 5.5 is Anthropic’s fast, efficient model for high-volume applications where responsiveness and cost matter. It is designed for summarization, classification, information extraction, request routing, and live customer support, making it well suited to everyday automation and interactive assistants. With a 1-million-token context window and adaptive thinking, Haiku 5.5 combines large-context processing with adjustable reasoning depth. It also works well as a specialized subagent, handling focused research, document lookups, and supporting coding tasks within larger workflows. Choose Haiku 5.5 for narrowly scoped work that needs to run quickly and economically at scale.

Total Context

1M

Max Output

128K

Released

Oct 7, 2026

Input$0.1 / M tokens
Output$0.5 / M tokens
Cache Read$0.01 / M tokens

Google

Gemini 4 Argon

gemini-4-argon

Gemini 4 Argon is Google's flagship reasoning model, built for complex software engineering, deep mathematical and scientific problem solving, and autonomous multi-step agents. Google trained it to keep context across long, multi-step tasks and to turn long documents mixing text, tables, charts and images into a single grounded output. It leads the Vals Index across finance, legal, coding and tax work, and supports long decode of up to 1M output tokens in a single response.

Total Context

1M

Max Output

65.5K

Released

Oct 13, 2026

Input$4 / M tokens
Output$20 / M tokens
Cache Read$0.2 / M tokens

Google

Gemini 4 Hydrogen

gemini-4-hydrogen

Gemini 4 Hydrogen is Google's high-efficiency workhorse in the Gemini 4 family, designed for high-throughput agentic workflows, coding, and multi-model routing at lower cost and latency than Gemini 4 Carbon. It accepts text, image, audio, video, and file input, supports long decoding of up to 1M output tokens via continuation, and controls thinking depth with low, medium, high, and xhigh reasoning effort.

Total Context

1M

Max Output

65.5K

Released

Nov 12, 2026

Input$1.5 / M tokens
Output$7.5 / M tokens
Cache Read$0.075 / M tokens

StepFun

Step 5 Preview

step-5-preview

Step 5 Preview is StepFun’s flagship model for complex reasoning, software engineering, and multi-step agent workflows. With native support for text, image, and video input and a 1-million-token context window, it can analyze extensive documents, codebases, charts, and visual content within a single request. Built for practical problem-solving, the model supports tool calling, structured outputs, and adjustable reasoning effort. Its strengths span coding, research, and professional knowledge work, with particular emphasis on financial analysis. Step 5 Preview is a strong choice for applications that need to combine information from multiple sources and work toward detailed, actionable results.

Total Context

1M

Max Output

1M

Released

Sep 16, 2026

Input$1 / M tokens
Output$2.7 / M tokens
Cache Read$0.05 / M tokens

OpenAI

GPT-6.1 Sol Pro

gpt-6.1-sol-pro

GPT-6.1 Sol Pro is an enhanced reasoning version of GPT-6.1 Sol. It uses the same underlying model as GPT-6.1 Sol but enables reasoning.mode=pro, allowing the model to spend significantly more computation on each request. The additional reasoning budget helps GPT-6.1 Sol Pro handle harder problems that require deeper analysis, longer planning, and higher accuracy. It is designed for tasks where getting the best possible answer is more important than response speed or cost efficiency. Compared with GPT-6.1 Sol, GPT-6.1 Sol Pro consumes significantly more reasoning tokens per request, resulting in higher usage costs and longer response times.

Total Context

1.1M

Max Output

128K

Released

Sep 29, 2026

Input$2 / M tokens
Output$10 / M tokens
Cache Read$0.1 / M tokens

OpenAI

GPT-6.1 Terra

gpt-6.1-terra

GPT-6.1 Terra is the mid-tier model in OpenAI's GPT-6.1 series, priced below GPT-6.1 Sol. It suits everyday agentic, coding, and writing work, excels at understanding screenshots and on-screen content, and is efficient for high-volume workloads.

Total Context

1.1M

Max Output

128K

Released

Oct 26, 2026

Input$1 / M tokens
Output$5 / M tokens
Cache Read$0.05 / M tokens

Anthropic

Claude Sonnet 5.5

claude-sonnet-5.5

Claude Sonnet 5.5 is a next-generation AI model designed for advanced reasoning, coding, analysis, and complex agentic workflows. It delivers strong performance across software development, long-context understanding, multimodal tasks, and professional knowledge work, offering a balanced combination of intelligence, speed, and efficiency. With enhanced reasoning capabilities and improved instruction following, Claude Sonnet 5.5 excels at writing and debugging code, analyzing complex documents, solving multi-step problems, and powering AI agents that require reliable decision-making and tool usage.

Total Context

1M

Max Output

128K

Released

Sep 28, 2026

Input$2 / M tokens
Output$10 / M tokens
Cache Read$0.2 / M tokens

OpenAI

GPT-6.1 Terra Pro

gpt-6.1-terra-pro

GPT-6.1 Terra Pro is the same underlying model as GPT-6.1 Terra, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Total Context

1.1M

Max Output

128K

Released

Oct 26, 2026

Input$1 / M tokens
Output$5 / M tokens
Cache Read$0.05 / M tokens

xAI

Grok 4.7

grok-4.7

Grok 4.7 is xAI's latest generation model, built for demanding coding tasks, complex reasoning, and knowledge-intensive applications. It improves long-horizon problem solving, self-verification, and workflow execution, making it a strong fit for developers building intelligent assistants and AI-powered products. Grok 4.7 delivers strong performance across programming, analysis, and general-purpose AI tasks while maintaining competitive efficiency for production deployments.

Total Context

500K

Max Output

500K

Released

Sep 21, 2026

Input$1.6 / M tokens
Output$4.8 / M tokens
Cache Read$0.4 / M tokens

Anthropic

Claude Opus 5.5

claude-opus-5.5

Claude Opus 5.5 is Anthropic's flagship intelligence model, designed for demanding workloads that require deep reasoning, reliability, and long-running task execution. It delivers strong performance in software engineering, agentic workflows, professional analysis, and knowledge-intensive tasks. With improved efficiency and enhanced safeguards, Claude Opus 5.5 is optimized for developers building sophisticated AI applications that require consistent performance across complex tasks.

Total Context

1M

Max Output

128K

Released

Sep 22, 2026

Input$4 / M tokens
Output$20 / M tokens
Cache Read$0.2 / M tokens

OpenAI

GPT-6 Sol

gpt-6-sol

GPT-6 Sol is a high-performance model in OpenAI's GPT-6 series, designed for complex reasoning tasks, software development, and professional AI workflows. It offers enhanced intelligence for multi-step problem solving, advanced coding, data analysis, and agent-based applications. GPT-6 Sol combines strong reasoning capability with efficient inference, making it suitable for developers who need frontier-level performance for production AI systems.

Total Context

1.1M

Max Output

128K

Released

Sep 22, 2026

Input$2 / M tokens
Output$10 / M tokens
Cache Read$0.2 / M tokens

Popular model recommendations

Start with high-signal models from the live catalog, then open a detail page to compare context, endpoints, and effective pricing.

OpenAI

GPT-6.2 Astra

GPT-6.2 Astra is OpenAI's flagship GPT-6.2 model, built for the hardest projects. It suits real-world engineering, professional work, and scientific research, with stronger visual reasoning for websites, slideshows, and 3D designs. OpenAI reports new eval highs at less than a third of GPT-6 Astra's cost per task.

Total Context

1.1M

Input Price

$8 / M tokens

View model

OpenAI

GPT-6.2 Astra Pro

GPT-6.2 Astra Pro uses the same underlying model as GPT-6.2 Astra, served with reasoning.mode set to pro for higher-quality responses on complex tasks. Cost note: pro mode uses far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6.2 Astra and takes much longer. Use it for hard, high-stakes problems where extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6.2 Astra instead.

Total Context

1.1M

Input Price

$8 / M tokens

View model

Tencent

Hy4 flash

hy4-flash has roughly one-quarter the parameters of the flagship model. With lower pricing, it matches flagship open-source performance on everyday tasks — well suited to coding assistance, agent workflows, document and chart understanding, and high-volume daily office work.

Total Context

1M

Input Price

$0.035 / M tokens

View model

Moonshot

Kimi K3.1 Pro

Kimi K3.1 Pro is Moonshot AI's flagship open-weight multimodal reasoning model and a post-trained successor to Kimi K3, featuring a 1M-token context window. It is built for long-horizon coding, 3D creation, financial research, and end-to-end agentic knowledge work, with native image and video understanding.

Total Context

1M

Input Price

$2 / M tokens

View model

Alibaba

Qwen4-Max

Qwen4-Max is Tongyi Qianwen's new-generation flagship multimodal model, combining text generation, image and video understanding, and deep reasoning. It is suited for coding, office work, front-end design, and complex agent tasks. The model supports web search, tool calling, and structured output, and is compatible with OpenAI and Anthropic APIs, helping developers build intelligent applications with multimodal understanding and task execution capabilities.

Total Context

1M

Input Price

$1.425 / M tokens

View model

OpenAI

GPT-6.2 Sol

GPT-6.2 Sol is an upgrade to OpenAI's GPT-6.1 Sol, positioned below the flagship GPT-6.2 Astra in the GPT-6.2 series. It is an efficient, versatile model for everyday work, suited to multistep agentic workflows and clear, natural writing across technical and non-technical tasks.

Total Context

1.1M

Input Price

$2 / M tokens

View model

Model Comparison

Quick comparison against selected catalog neighbors.

Model catalog FAQ

A quick guide for choosing, comparing, and using models from the TokenHub catalog.

How should I choose a model from this list?

+

Start with your workload. Use the filters to narrow by provider, tags, endpoint type, and billing group, then compare context size, output limit, modalities, and input or output pricing.

What does effective price mean?

+

Effective price applies the active billing group ratio to the model pricing data. It helps you estimate the real input, output, or per-request cost for the group you are using.

Can I use these models through API endpoints?

+

Yes. Open a model detail page to see the supported endpoint types and documentation links. Availability can differ by model, provider, and current routing configuration.

Why do context window and max output matter?

+

The context window controls how much prompt and conversation history a model can read. Max output controls how much text it can generate in one response, which matters for long-form writing, coding, and document tasks.