LLM API pricing, tracked from launch.

o1-pro is OpenAI's reasoning model trained with reinforcement learning to perform extended chain-of-thought processing on complex problems, with a 200k token context window.

o1-pro is OpenAI's reasoning model trained with reinforcement learning to perform extended chain-of-thought processing on complex problems.

OpenAI's most advanced frontier model from the GPT-5.4 series with a 1.05M token context window for handling large documents and complex tasks.

Highest-accuracy reasoning variant of GPT-5.5.

OpenAI's frontier-grade reasoning model with a 400k token context window, released in December 2025 as part of the GPT-5.2 generation.

GPT-5 Pro is OpenAI's advanced general-purpose reasoning model with a 400k token context window, released in October 2025 as the predecessor to GPT-5.1.

OpenAI's most advanced general-purpose model from the GPT-5.4 series, optimized for batch processing with a 1.05M token context window.

OpenAI's highest-accuracy reasoning variant from the GPT-5.5 series, optimized for batch processing with a 1M+ token context window.

OpenAI's frontier-grade reasoning model with a 400k token context window, optimized for batch processing and released in December 2025 as part of the GPT-5.2 generation.

Premium high-compute reasoning variant launched alongside the 80% o3 price cut — 87% cheaper than the o1-pro it replaced.

o3 Pro is OpenAI's high-compute reasoning variant in the o-series, trained with reinforcement learning to perform extended chain-of-thought processing on complex problems.

GPT-5 Pro (batch) is OpenAI's advanced general-purpose reasoning model optimized for asynchronous batch processing, with a 400k token context window.

GPT-6 Astra is OpenAI's flagship model for complex end-to-end work, positioned as the successor to the GPT-5.6 series with an expanded 1.05M token context window.

Claude Fable 5.1 is Anthropic's most capable reasoning model, an incremental improvement over Claude Fable 5 with a 1M token context window.

Anthropic's most powerful model — a new Mythos-class tier above Opus, aimed at the hardest reasoning and agentic work.

OpenAI's high-compute reasoning model in the o-series, designed for extended chain-of-thought processing on complex problems, available only for batch processing.

o3 Deep Research is OpenAI's specialized model for extended reasoning and in-depth analysis tasks, featuring a 200k token context window.

Claude Opus 4.1 in batch mode is Anthropic's flagship reasoning model optimized for asynchronous, non-real-time workloads with a 200,000-token context window.

Fugu Ultra v2 is a multi-agent orchestration model that routes requests across specialized language models rather than relying on a single monolithic architecture.

OpenAI's reasoning model trained with reinforcement learning to perform extended chain-of-thought processing, optimized for asynchronous batch processing.

A large-context model from Sakana with a 1-million-token window for processing extended documents and codebases.

GPT-4 Turbo Preview is OpenAI's intermediate-tier large language model with vision capabilities, a 128K token context window, and support for JSON mode and function calling.

OpenAI's frontier model for complex professional workloads. 1M token context, text + image input, native computer use.

OpenAI's flagship model for complex end-to-end work, run in batch mode with pro-level reasoning for higher-quality responses on difficult tasks.

OpenAI's flagship model released in 2026, designed for complex end-to-end work with a 1.05M token context window, available in batch mode.

Claude Fable 5.1 is Anthropic's most capable reasoning model in batch mode, an incremental improvement over Claude Fable 5 with a 1 million token context window.

Claude Fable 5 in batch mode is Anthropic's most capable reasoning model, optimized for asynchronous processing with a 1 million token context window.

Anthropic's flagship reasoning model designed for complex multi-step problems, software development, and extended agentic work.

Introduced the 67% Opus price cut — from $15/$75 (Opus 4/4.1) to $5/$25 — which held through all subsequent Opus releases.

First Opus with a 1M context window at standard pricing. Same $5/$25 as Opus 4.5.

Previous-generation Opus. Highly autonomous; strong on agentic, vision and memory tasks.

Most capable Opus-tier model: highly autonomous, strong on long-horizon agentic work and knowledge tasks. 1M context at standard pricing.

A 2.8T parameter multimodal reasoning model from Moonshot AI with a 1M+ token context window, available as a batch processing variant.

GPT-4 Turbo is OpenAI's intermediate-tier language model with vision capabilities, a 128K token context window, and support for JSON mode and function calling.

OpenAI's general-purpose frontier model optimized for batch processing, with a 1M+ token context window for handling large documents and complex reasoning.

GPT-4o is OpenAI's multimodal large language model that processes text and images with a 128K token context window, released in May 2024.

Sonar Pro is a large language model from Perplexity with a 200k token context window that integrates web search capabilities.

OpenAI's general-purpose frontier model with a 1.05M token context window for handling large documents and complex reasoning tasks.

Previous Sonnet generation. Same $3 / $15 pricing as 4.6 — Sonnet pricing has held flat.

Anthropic's best balance of speed and intelligence. 1M context window. Sonnet pricing has held flat at $3/$15 across all 4.x releases.

GPT-5.2 is OpenAI's general-purpose reasoning model with a 400k token context window, released in December 2025 as the successor to GPT-5.1.

GPT-5.2-Codex is OpenAI's coding-specialized model from the GPT-5.2 generation, built for software development tasks with a 400k token context window.

GPT-5.3-Codex is OpenAI's coding-specialized model from the GPT-5.3 generation, built for software development tasks with a 400k token context window.

GPT-5.3 Chat is OpenAI's general-purpose conversational model with a 128k token context window, positioned between the older GPT-5.2 generation and the newer GPT-5.4 frontier series.

Claude Opus 4.5 in batch mode is Anthropic's flagship reasoning model optimized for asynchronous, non-real-time workloads with a 200k-token context window.

Claude Opus 4.6 in batch mode, Anthropic's flagship reasoning model optimized for asynchronous processing with a 1 million token context window.

Claude Opus 4.7 in batch mode is Anthropic's previous-generation flagship reasoning model optimized for asynchronous processing, with a 1 million token context window.

Claude Opus 4.8 in batch mode is Anthropic's previous-generation flagship reasoning model optimized for asynchronous processing, with a 1 million token context window.

Claude Opus 5 in batch mode, Anthropic's flagship reasoning model optimized for asynchronous processing with a 1M token context window.

Nova Premier 1.0 is Amazon's flagship general-purpose language model with a 1-million-token context window.

GPT-5.6 Terra is a midrange model in OpenAI's GPT-5.6 series that balances capability and efficiency with a 1.05M token context window.

Previous Gemini Pro generation with the same context-tiered pricing model. Superseded by Gemini 3.1 Pro.

Google's frontier model. Context-tiered pricing: $2/$12 up to 200K tokens, $4/$18 beyond. Figures shown are the ≤200K tier.

Kimi K3 is a multimodal reasoning model from Moonshot AI with a 1M+ token context window designed for complex problem-solving and long-context tasks.

OpenAI's flagship reasoning model in the GPT-5.6 series with a 1.05M token context window.

Claude Sonnet 5 is Anthropic's mid-tier model with a 1 million token context window and adaptive thinking capabilities selectable at multiple reasoning effort levels.

Cohere's enterprise RAG flagship, marketed as Command R+. The August 2024 refresh cut pricing from $3/$15 to $2.50/$10.

GPT-4o (August 2024 release) is OpenAI's multimodal large language model that processes text and images with a 128K token context window.

Cohere's August 2024 update to Command R+, a large language model with a 128K token context window optimized for production workloads.

Inflection 3 Pi is a conversational model that powers the Pi chatbot, designed for dialogue with built-in knowledge of recent events.

Inflection 3 Productivity is a compact instruction-following model optimized for structured outputs and rule-based tasks with access to recent information.

A multimodal large language model from OpenAI released in November 2024 that processes text and images with a 128K token context window.

GPT-4o with integrated web search capability, allowing retrieval and citation of current information from the internet during inference.

Cohere's large language model with a 256K token context window, extending the depth available in the Command series for handling longer documents and extended conversations.

GPT-5 Chat is OpenAI's general-purpose conversational model with a 128k token context window, positioned as a mid-range option within the GPT-5 family.

GPT-5 Codex is OpenAI's code-specialized model optimized for software engineering and programming tasks, built on the GPT-5 foundation with a 400k token context window.

A coding-specialized model from OpenAI's GPT-5.1 generation with a 400k token context window.

GPT-5.1 is OpenAI's general-purpose reasoning model with a 400k token context window, released in November 2025 as the predecessor to GPT-5.2.

OpenAI's coding-specialized model from the GPT-5.1 generation with a 400k token context window, released before the December 2025 GPT-5.2 generation.

A multimodal model from OpenAI that processes text, images, and audio inputs with a 128k token context window.

First Gemini with native thinking/reasoning. Context-tiered: $1.25/$10 ≤200K, $2.50/$15 beyond.

Launched at commodity pricing ($1.25/$10) so low TechCrunch said it may spark a price war. Superseded by GPT-5.5.

Latest Flash model — 3× pricier than Gemini 3 Flash ($0.50/$3), signaling the end of ultra-cheap Flash pricing.

Aion-1.0 is a large language model from AionLabs with a 131k-token context window, released before the company's shift toward multi-model systems.

A research-focused model from Perplexity that autonomously searches, retrieves, and synthesizes information across complex topics with a 128k token context window.

A reasoning model based on DeepSeek R1 that uses chain-of-thought processing to work through complex problems step-by-step, with access to Perplexity's search integration.

Jamba Large 1.7 is a hybrid architecture model from AI21 that combines Transformers with Mamba state-space layers to handle extended context efficiently.

o4 Mini Deep Research is OpenAI's lightweight model for extended reasoning and research tasks with a 200k token context window.

Workhorse text model with 1M context. Positioned between the budget GPT-5 and premium GPT-5.5.

OpenAI's frontier reasoning model. Received an 80% price cut in June 2025 alongside the o3-pro launch.

Claude Sonnet 4.6 in batch mode is Anthropic's speed-focused reasoning model optimized for asynchronous processing with a 1 million token context window.

Claude Sonnet 4.5 in batch mode is Anthropic's mid-tier model optimized for asynchronous, non-real-time workloads with a 1 million token context window.

GPT-5.4 (batch) is OpenAI's general-purpose frontier model optimized for batch processing, with a 1.05M token context window for handling large documents and complex reasoning tasks.

Alibaba's latest flagship Qwen model with 1M context, announced at Alibaba Cloud Summit.

Consolidated model folding Medium 3.1, Magistral reasoning, and Devstral 2 into one set of weights with a per-request reasoning toggle.

GPT-5.2 (batch) is OpenAI's general-purpose reasoning model with a 400k token context window, optimized for asynchronous batch processing workflows.

Qwen3.6 Max Preview is Alibaba's preview-stage flagship model from the Qwen3.6 generation with a 262K token context window.

Fugu Max is a multi-agent orchestration model that routes requests across specialized language models to balance performance and efficiency.

Qwen3.8 2.4T A95B is an open-weight sparse mixture-of-experts model from Alibaba with 95 billion active parameters from a 2.4 trillion total parameter pool, supporting text, image, and video inputs.

GPT-5.6 Terra Pro (batch) is OpenAI's pro-level midrange model in the GPT-5.6 series, optimized for batch processing with a 1.05M token context window.

xAI's frontier-tier model with a 500k token context window, released after Grok 4.5.

Qwen3.8 2.4T A95B is Alibaba's large multimodal model from the Qwen3.8 generation with support for text, image, and video inputs across a 1 million token context window.

GPT-5.6 Terra is a midrange model in OpenAI's GPT-5.6 series that balances capability and efficiency, available in batch mode with a 1.05M token context window.

Qwen3.8 Max is Alibaba's flagship language model from the Qwen3.8 generation, released in August 2026, with a 1 million token context window.

Google's frontier model optimized for batch processing, supporting the full 1M token context window with the same capability tier as Gemini 3.1 Pro.

Grok 4.5 is xAI's intermediate reasoning model with a 500k token context window, positioned between the coding-focused Grok Build 0.1 and the frontier Grok 4.6.

A multi-model system from AionLabs that combines specialized models for collaborative generation, designed around roleplaying and storytelling tasks.

Mistral's instruction-tuned mixture-of-experts model with 39B active parameters selected from 141B total, designed for efficient inference across diverse tasks.

Mistral's flagship model with 131k token context window, designed for reasoning, code generation, JSON formatting, and conversational tasks.

Palmyra X5 is Writer's large language model with a 1 million token context window.

GPT-4o (batch) is OpenAI's multimodal large language model optimized for asynchronous batch processing, handling text and images with a 128K token context window.

GPT-5 Codex is OpenAI's code-specialized model optimized for software engineering and programming tasks, built on the GPT-5 foundation with a 400k token context window.

OpenAI's flagship reasoning model in the GPT-5.6 series, available in batch mode with a 1.05M token context window.

GPT-5.6 Sol Pro (batch) is OpenAI's flagship reasoning model in the GPT-5.6 series, served in batch mode with pro-level reasoning for higher-quality responses on complex tasks and a 1.05M token context window.

Google's Pro reasoning model from the Gemini 2.5 family with native thinking capabilities, optimized for batch processing with a 1-million-token context window.

OpenAI's general-purpose model with a 400k token context window, optimized for batch processing where latency is not a constraint.

GPT-5.1 (batch) is OpenAI's general-purpose reasoning model with a 400k token context window, optimized for asynchronous batch processing workflows.

Claude Sonnet 5 is Anthropic's mid-tier model optimized for batch processing, offering adaptive thinking capabilities with a 1 million token context window.

This is a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet(https://openrouter.ai/anthropic/claude-3.5-sonnet) and Opus(https://openrouter.ai/anthropic/claude-3-opus). The model is fine-tuned on top of [Qwen2.5 72B](https://openrouter.ai/qwen/qwen-2.5-72b-instruct).

Fastest and most cost-effective Claude model. Pricing held flat since launch — no newer Haiku has been released.

Google's text-only model optimized for batch processing with a 1M token context window.

A lightweight variant of GPT-5.4 with a 400k token context window, optimized for speed and efficiency in high-volume workloads.

GLM 5.3 is Z.ai's latest foundation model with a 1M-token context window for processing extended documents and complex reasoning tasks.

o3 Mini High is OpenAI's cost-efficient reasoning model with extended chain-of-thought processing, multimodal input support, and a 200k token context window.

o4-mini with reasoning_effort set to high, a compact reasoning model that performs extended chain-of-thought processing with multimodal input support.

Cost-effective reasoning model with multimodal support. Replaced o3-mini at the same price point.

Meta's multimodal reasoning model designed for long-running agentic, multi-agent, and coding workflows with a 1-million-token context window.

Muse Spark 1.2 is Meta's multimodal reasoning model that processes text and images together with a 1-million-token context window.

Meta's first paid model API — a multimodal reasoning model built for agentic tasks, tool/computer use, coding, and multimodal understanding. Meta AI/Superintelligence Labs' answer to Anthropic/OpenAI's agentic-coding models.

Open-weight multimodal mixture-of-experts model with 41B active parameters from 975B total and a 524K token context window.

An open-weight multimodal mixture-of-experts model with 12B active parameters from 276B total parameters and a 524K token context window.

A Japanese-specialized reasoning model from Sakana AI, built on Kimi K2.6 with additional training for Japanese language and business contexts.

GPT-4.1 (batch) is OpenAI's text model optimized for asynchronous batch processing, with a 1 million token context window.

OpenAI's frontier reasoning model with 200k token context, optimized for batch processing of complex problems requiring extended chain-of-thought reasoning.

A coding-specialized model from Moonshot AI's Kimi K2 family with a 262K token context window and mixture-of-experts architecture.

GLM 5V Turbo is Z.ai's optimized multimodal inference variant with support for text, image, and video inputs alongside a 202K-token context window.

GPT-3.5 Turbo 16k is OpenAI's lightweight language model for chat and code tasks with a 16K token context window.

A 235-billion-parameter multimodal vision-language model from Alibaba that processes text, images, and video with extended reasoning capabilities.

GLM 5 Turbo is Z.ai's optimized inference variant of their GLM 5 foundation model, offering faster processing with a 202K-token context window.

Moonshot AI's latest open-weight frontier model with 300-agent swarm support.

Qwen3 Max Thinking is Alibaba's large language model with extended reasoning capabilities and a 262K token context window.

Gemini 3.8 Flash is Google's latest multimodal Flash model with a 1M token context window, supporting text, image, video, and audio inputs.

Mistral Medium 3.5 is a 128B dense model that consolidates instruction-following, extended reasoning, and code generation capabilities into a single set of weights with optional per-request reasoning.

Google's multimodal Flash model supporting text, image, video, and audio inputs with a 1M token context window.

Google's text-only multimodal Flash model with a 1M token context window, positioned between the lightweight Gemini 3.5 Flash Lite and the multimodal Gemini 3.7 Flash.

A 55-billion-parameter mixture-of-experts model from NVIDIA with a hybrid Transformer-Mamba architecture and 512K token context window.

Kimi K2.7 Code is a specialized language model from Moonshot AI with a 262K token context window optimized for software development tasks.

Qwen3.5 397B A17B is a 397-billion-parameter text language model from Alibaba's Qwen3.5 generation with a 262K token context window.

A routing model that directs requests to other LLMs based on task characteristics and cost-performance tradeoffs.

Qwen3 Coder Plus is Alibaba's code-focused large language model with a 1-million-token context window, positioned as the higher-capacity tier in the Qwen3 Coder lineup.

A multimodal model from Amazon designed to balance accuracy and speed across a broad range of tasks, with a 300K-token context window.

A 55-billion-parameter mixture-of-experts model from NVIDIA with a hybrid Transformer-Mamba architecture and 512K token context window.

GLM 5.1 is Z.ai's foundation model with a 202K-token context window for processing extended documents and conversations.

Qwen3.8 27B is a 27-billion-parameter text language model from Alibaba's Qwen3.8 generation with a 1 million token context window.

Seed-2.0-Code is a specialized coding model from ByteDance Seed with a 262k token context window, designed for software development tasks.

This is [Sao10K](/sao10k)'s experiment over [Euryale v2.2](/sao10k/l3.1-euryale-70b).

Hermes 4 405B is a 405-billion parameter instruction-tuned language model from Nous with a 131K token context window.

Relace Search is a specialized model designed to locate and extract relevant code segments and documentation from large codebases and repositories.

Multimodal model with Agent Swarm technology. Superseded by K2.6.

Cost-effective Flash model. Superseded by Gemini 3.5 Flash which tripled the price.

A coding-focused language model from Kwaipilot with a 256K token context window.

Hy4 preview is a mixture-of-experts model from Tencent with 49B active parameters and a 1M-token context window, designed for coding, tool use, and multi-step reasoning tasks.

Seed 2.1 Turbo is a multimodal model from ByteDance Seed designed for coding tasks and agentic workflows, with a 262k token context window.

Claude Haiku 4.5 in batch mode is Anthropic's fastest and most cost-effective model, optimized for asynchronous processing with a 200k-token context window.

A lightweight text-only model from Google with a 1M token context window, released in mid-2026 as a faster, more efficient alternative to the standard Flash tier.

DeepSeek R1 is an open-weight reasoning model that applies chain-of-thought inference to work through complex problems step-by-step before answering.

Kimi K2 0905 is a Mixture-of-Experts language model from Moonshot AI with a 262K token context window, released in September 2025.

Kimi K2 Thinking is a reasoning-focused model from Moonshot AI that uses extended chain-of-thought to work through complex problems before generating responses.

Nova 2 Lite is Amazon's lightweight multimodal model that processes text, images, and video with a 1-million-token context window.

xAI's multi-agent reasoning model with a 2 million token context window, released before Grok 4.3.

Predecessor to Grok 4.3 with 2M context. Later aligned to Grok 4.3 pricing.

xAI's current flagship — aggressively priced for a frontier model, with a 1M token context window.

First Flash model with thinking. Flat pricing (no context tiers). Superseded by Gemini 3 Flash.

A 30.5-billion-parameter mixture-of-experts model from Alibaba that activates 3.3 billion parameters per token and uses extended reasoning capabilities.

A 30-billion-parameter multimodal vision-language model from Alibaba that processes text, images, and video with extended reasoning capabilities and a 131k-token context window.