Model Wiki

GLM-4.7 is a glm model from Zhipu AI (~204.8K context, input around $0.6/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-4.6V-Flash is a glm-flash model from Zhipu AI (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-4.6V is a glm model from Zhipu AI (~128K context, input around $0.3/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-4.6 is a glm model from Zhipu AI (~204.8K context, input around $0.6/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-4.5V is a glm model from Zhipu AI (~64K context, input around $0.6/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-4.5-Flash is a glm-flash model from Zhipu AI (~131.1K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-4.5 is a glm model from Zhipu AI (~131.1K context, input around $0.6/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-4.5-Air is a glm-air model from Zhipu AI (~131.1K context, input around $0.2/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

MiniMax M3.1 Flash Preview supports multimodal input, a 1M context window and tunable thinking depth, currently only through M Plan and MiniMax Code.

MiniMax M3 is a MiniMax model for coding, agent workflows and complex production tasks, often evaluated for Chinese assistants, generation and multimedia workflows.

MiniMax Speech 01 focuses on voice generation and multimedia experiences rather than general text conversation.

MiniMax M1 is a MiniMax model for long-context reasoning, coding and complex task planning, often evaluated for Chinese assistants, generation and multimedia workflows.

MiniMax Text 01 is a MiniMax model for general language understanding, generation and agent workflows, often evaluated for Chinese assistants, generation and multimedia workflows.

ABAB6.5 Chat is a MiniMax model for general conversation, long-text understanding and complex interactions, often evaluated for Chinese assistants, generation and multimedia workflows.

ABAB6.5s Chat is a MiniMax model for Chinese chat, writing and business assistant workloads, often evaluated for Chinese assistants, generation and multimedia workflows.

abab7-chat is a large language model from MiniMax. This entry summarizes its positioning, typical use cases, strengths, limitations and related pricing signals for quick comparison.

abab6.5s-chat is a large language model from MiniMax. This entry summarizes its positioning, typical use cases, strengths, limitations and related pricing signals for quick comparison.

MiniMax-M2.5-highspeed is a minimax model from MiniMax (~204.8K context, input around $0.6/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

MiniMax-M2.5 is a minimax model from MiniMax (~204.8K context, input around $0.3/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

MiniMax-M2.1 is a minimax model from MiniMax (~204.8K context, input around $0.3/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

MiniMax-M2 is a minimax model from MiniMax (~196.6K context, input around $0.3/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Step 3.7 Flash is a StepFun model for efficient reasoning, complex Chinese tasks and production assistants, commonly evaluated for Chinese assistants, document and multimodal workflows.

Step 1V 8K is a StepFun model for visual Q&A, multimodal analysis and image-text understanding, commonly evaluated for Chinese assistants, document and multimodal workflows.

Step 2 Mini is a StepFun model for low-latency and cost-sensitive high-volume workloads, commonly evaluated for Chinese assistants, document and multimodal workflows.

Step 1 128K is a StepFun model for very long-document analysis and retrieval-augmented workflows, commonly evaluated for Chinese assistants, document and multimodal workflows.

Step 3.5 Flash is a StepFun model for reasoning, Q&A and business assistant workloads, commonly evaluated for Chinese assistants, document and multimodal workflows.

Step 1 32K is a StepFun model for document understanding, summarization and knowledge Q&A, commonly evaluated for Chinese assistants, document and multimodal workflows.

Step-1 8K is a large language model from StepFun. This entry summarizes its positioning, typical use cases, strengths, limitations and related pricing signals for quick comparison.

Step-2 16K is a large language model from StepFun. This entry summarizes its positioning, typical use cases, strengths, limitations and related pricing signals for quick comparison.

StepAudio 2.5 ASR is a step model from stepfun. Suitable for assistants, content generation, knowledge Q&A and business automation

StepAudio 2.5 TTS is a step model from stepfun. Suitable for assistants, content generation, knowledge Q&A and business automation

Step 3.5 Flash 2603 is a AI model from stepfun (~256K context, input around $0.1/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Step TTS 2 is a step model from stepfun. Suitable for assistants, content generation, knowledge Q&A and business automation

Doubao Seed 2.1 Lite 260915 supports text generation, image/video understanding and audio understanding, deep thinking, GUI tasks, tool use and structured output.

Doubao Seed 2.1 Pro 260915 supports text generation and image/video understanding, deep thinking, GUI tasks, tool use and structured output.

GLM-5.3-Flash is a glm-flash model from Volcengine (~1M context, input around $0.11875/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

DeepSeek V4 Pro 0813 is a deepseek-thinking model from Volcengine (~1M context, input around $1.3359/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

DeepSeek V4 Flash 0731 is a deepseek-flash model from Volcengine (~1M context, input around $0.4453/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed Character is a seed model from Volcengine (~256K context, input around $0.11875/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 2.1 Pro is a seed model from Volcengine (~256K context, input around $0.8906/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-5.2 is a glm model from Volcengine (~1M context, input around $1.18747/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Doubao Vision Pro is a Volcengine Doubao model for image-text analysis, multimodal Q&A and visual content understanding, often evaluated for Chinese enterprise and multimodal workflows.

Doubao Seed 1.6 Thinking is a Volcengine Doubao model for complex reasoning, multi-step analysis and coding assistance, often evaluated for Chinese enterprise and multimodal workflows.

Doubao Seed 1.6 is a Volcengine Doubao model for general chat, reasoning and agent workflow evaluation, often evaluated for Chinese enterprise and multimodal workflows.

Doubao Lite is a Volcengine Doubao model for low-cost high-concurrency chat and lightweight text tasks, often evaluated for Chinese enterprise and multimodal workflows.

Doubao Pro is a Volcengine Doubao model for higher-quality Chinese assistants, content generation and business automation, often evaluated for Chinese enterprise and multimodal workflows.

Doubao Pro 32K is a large language model from Volcengine. This entry summarizes its positioning, typical use cases, strengths, limitations and related pricing signals for quick comparison.

Doubao Lite 32K is a large language model from Volcengine. This entry summarizes its positioning, typical use cases, strengths, limitations and related pricing signals for quick comparison.

Seed 2.0 Code is a seed model from Volcengine (~262.1K context, input around $0.47499/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 2.0 Mini is a seed model from Volcengine (~256K context, input around $0.02969/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 2.0 Lite is a seed model from Volcengine (~256K context, input around $0.08906/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 2.0 Pro is a seed model from Volcengine (~256K context, input around $0.47499/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 1.8 is a seed model from Volcengine (~256K context, input around $0.11875/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 1.6 is a seed model from Volcengine (~256K context, input around $0.11875/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 1.6 Flash is a seed model from Volcengine (~256K context, input around $0.02227/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Seed 1.6 Vision is a seed model from Volcengine (~256K context, input around $0.11875/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

MiMo V2.6 Pro UltraSpeed is a Xiaomi MiMo model for high-speed complex reasoning and coding workloads, useful for Chinese assistants, tool use and ecosystem-oriented applications.

MiMo V2.6 Flash is a Xiaomi MiMo model for low-cost high-volume chat, batch processing and quick responses, useful for Chinese assistants, tool use and ecosystem-oriented applications.

MiMo V2.6 Pro is a Xiaomi MiMo model for complex reasoning, coding and long-text tasks, useful for Chinese assistants, tool use and ecosystem-oriented applications.

MiMo-V2.5-Pro-UltraSpeed is a mimo model from Xiaomi (~1M context, input around $1.305/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

MiMo V2 Pro is a Xiaomi MiMo model for complex tasks, coding assistance and business automation, useful for Chinese assistants, tool use and ecosystem-oriented applications.

MiMo V2 Omni is a Xiaomi MiMo model for multimodal understanding and integrated interactive experiences, useful for Chinese assistants, tool use and ecosystem-oriented applications.

MiMo V2 Flash is a Xiaomi MiMo model for low-latency high-frequency chat and quick responses, useful for Chinese assistants, tool use and ecosystem-oriented applications.

MiMo V2.5 Pro is a Xiaomi MiMo model for complex reasoning, coding and long-text tasks, useful for Chinese assistants, tool use and ecosystem-oriented applications.

MiMo V2.5 is a Xiaomi MiMo model for Chinese conversation, content generation and tool-use workflows, useful for Chinese assistants, tool use and ecosystem-oriented applications.

ERNIE 5.0 is listed in Baidu Qianfan’s flagship catalog for text generation, visual understanding and deep thinking.

ERNIE 5.1 is listed in Baidu Qianfan’s flagship catalog for text generation.

ERNIE Lite is a Baidu ERNIE/Qianfan model for cost-sensitive Q&A, summarization and content generation, commonly evaluated for Chinese enterprise workloads.

ERNIE Speed is a Baidu ERNIE/Qianfan model for low-latency conversations and high-frequency basic text tasks, commonly evaluated for Chinese enterprise workloads.

ERNIE X1 is a Baidu ERNIE/Qianfan model for complex analysis, logical reasoning and multi-step problem solving, commonly evaluated for Chinese enterprise workloads.

ERNIE 4.5 Turbo is a Baidu ERNIE/Qianfan model for multi-scenario Chinese tasks, knowledge Q&A and business assistants, commonly evaluated for Chinese enterprise workloads.

ERNIE 4.0 Turbo 8K is a Baidu ERNIE/Qianfan model for Chinese understanding, content generation and enterprise applications, commonly evaluated for Chinese enterprise workloads.

Baidu (ERNIE)ernie

ERNIE 4.0 Turbo 8K is a large language model from Baidu. This entry summarizes its positioning, typical use cases, strengths, limitations and related pricing signals for quick comparison.

GLM-5.3 is a glm model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

GLM-5.2 is a glm model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mixtral 8x7B is a Mistral model for general language tasks, research and self-hosted deployments, with options across enterprise, coding, vision and open-weight workflows.

Pixtral Large is a Mistral model for image-text understanding, visual Q&A and multimodal analysis, with options across enterprise, coding, vision and open-weight workflows.

Ministral 8B is a Mistral model for edge deployment, low-cost usage and basic text tasks, with options across enterprise, coding, vision and open-weight workflows.

Mistral Large is a Mistral model for complex reasoning, enterprise Q&A and multilingual tasks, with options across enterprise, coding, vision and open-weight workflows.

Mistral Small is a Mistral model for low-latency high-concurrency and cost-sensitive workloads, with options across enterprise, coding, vision and open-weight workflows.

Codestral is a Mistral model for code completion, code generation and developer assistance, with options across enterprise, coding, vision and open-weight workflows.

Mistral Medium 3.5 is a mistral-medium model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Medium (latest) is a mistral-medium model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Small 4 is a mistral-small model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Voxtral Mini TTS (latest) is a voxtral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Voxtral Mini (latest) is a voxtral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Devstral 2 is a devstral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Devstral Small 2 is a devstral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Devstral 2 is a devstral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Devstral 2 (latest) is a devstral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Medium 3.1 is a mistral-medium model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Voxtral Small (latest) is a voxtral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Devstral Small is a devstral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Devstral Medium is a devstral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Small 3.2 is a mistral-small model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Medium 3 is a mistral-medium model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Devstral Small 2505 is a devstral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Magistral Small is a magistral-small model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Magistral Medium (latest) is a magistral-medium model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Large 2.1 is a mistral-large model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Large 3 is a mistral-large model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Ministral 3B (latest) is a ministral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Pixtral 12B is a pixtral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Open Mistral Nemo is a mistral-nemo model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Nemo is a mistral-nemo model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mixtral 8x22B is a mixtral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral Embed is a mistral-embed model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mixtral 8x7B is a mixtral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Mistral 7B is a mistral model from Mistral. Suitable for assistants, content generation, knowledge Q&A and business automation

Embed v5.0 Fast embeds text, images and mixed documents with 128k context. Pro prioritizes quality; Fast prioritizes lower latency. It returns vectors, not chat answers.

Embed v5.0 Pro embeds text, images and mixed documents with 128k context. Pro prioritizes quality; Fast prioritizes lower latency. It returns vectors, not chat answers.

North Small Translate 1.0 is a text MoE model purpose-built for machine translation across more than 50 languages through the Chat endpoint.

North Small Translate is a north model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

North Mini Code is a north model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Rerank v3.5 is a Cohere reranking model for search reranking, RAG refinement and answer-quality improvements.

Embed v4.0 is a Cohere embedding model for semantic search, RAG indexing, clustering and similarity workflows.

Command Light is a Cohere generation model for enterprise Q&A, RAG, agents and multilingual generation.

Command R is a Cohere generation model for enterprise Q&A, RAG, agents and multilingual generation.

Command R+ is a Cohere generation model for enterprise Q&A, RAG, agents and multilingual generation.

Command A is a Cohere generation model for enterprise Q&A, RAG, agents and multilingual generation.

Command A Plus is a command-a model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Tiny Aya Water is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Tiny Aya Global is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Tiny Aya Earth is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Tiny Aya Fire is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command A Translate is a command-a model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command A Reasoning is a command-a model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command A Vision is a command-a model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command A is a command-a model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Aya Vision 8B is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Aya Vision 32B is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command R7B Arabic is a command-r model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command R7B is a command-r model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Aya Expanse 32B is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Aya Expanse 8B is a AI model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command R+ is a command-r model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Command R is a command-r model from Cohere. Suitable for assistants, content generation, knowledge Q&A and business automation

Llama 4 Maverick is a Meta Llama model for complex multimodal tasks, agents and high-quality generation, commonly evaluated for open ecosystems, self-hosting and customizable AI products.

Llama 4 Scout is a Meta Llama model for multimodal, long-context and efficient reasoning workflows, commonly evaluated for open ecosystems, self-hosting and customizable AI products.

Llama 3.2 Vision is a Meta Llama model for image-text understanding, visual Q&A and multimodal applications, commonly evaluated for open ecosystems, self-hosting and customizable AI products.

Llama 3.1 405B is a Meta Llama model for complex reasoning, multilingual tasks and high-quality generation, commonly evaluated for open ecosystems, self-hosting and customizable AI products.

Llama 3.1 70B is a Meta Llama model for general language understanding, generation and enterprise self-hosting, commonly evaluated for open ecosystems, self-hosting and customizable AI products.

Llama 3.1 8B is a Meta Llama model for local deployment, basic chat and low-cost inference, commonly evaluated for open ecosystems, self-hosting and customizable AI products.

Groq-Llama-4-Maverick-17B-128E-Instruct is a llama model from llama (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Cerebras-Llama-4-Maverick-17B-128E-Instruct is a llama model from llama (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Llama-4-Scout-17B-16E-Instruct-FP8 is a llama model from llama (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Cerebras-Llama-4-Scout-17B-16E-Instruct is a llama model from llama (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Llama-4-Maverick-17B-128E-Instruct-FP8 is a llama model from llama (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Llama-3.3-70B-Instruct is a llama model from llama (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation

Llama-3.3-8B-Instruct is a llama model from llama (~128K context, input around $0/1M tokens). Suitable for assistants, content generation, knowledge Q&A and business automation