LLM Pricing: Top 15+ Providers Compared

LLM pricing now splits into two purchases: a per-token API budget for code that calls models programmatically, and an $8–$200 monthly subscription for everyday chat work.

• See major LLM subscription plans
• See LLMs ranked by performance, then enter your volume needs in tokens to see the exact pricing.

Hover over model names to view their benchmark results, real-world latency, and pricing, to assess each model’s efficiency and cost-effectiveness.

Ranking: Models are ranked by their average position across all benchmarks.

You can check the hallucination rates and reasoning performance of top LLMs in our benchmarks.

Understanding LLM pricing

Tokens: The Fundamental Unit of Pricing

Figure 1: Example of tokenization using the GPT-4o & GPT-4o mini tokenizer for the sentence “Identify New Technologies, Accelerate Your Enterprise.”¹

While providers offer a variety of pricing structures, per-token pricing is the most common. Tokenization methods differ across models; examples include:

• Byte-Pair Encoding (BPE): Splits words into frequent subword units, balancing vocabulary size and efficiency.²
  • Example: “unbelievable” → [“un”, “believ”, “able”]
• WordPiece: Similar to BPE but optimizes for language model likelihood, used in BERT.³
  • Example: “tokenization” → [“token”, “##ization”]. “token” is a standalone word; “##ization” is a suffix.
• SentencePiece: Tokenizes text without relying on spaces, effective for multilingual models like T5.⁴
  • Example: “natural language” → [” natural”, ” lan”, “guage”] or [” natu”, “ral”, ” language”].

Please note that the exact subwords depend on the training data and BPE/WordPiece process. To better understand these tokenization methods, watch the video below:

After grasping tokenization, an average price can be estimated based on the project token length. Table 2 outlines token ranges by content type, including UI prompts, email snippets, marketing blogs, detailed reports, and research papers, and notes that token counts vary across models. Once a model is chosen, its tokenizer can be used to estimate the average token count for the content.

Table 2: Typical content types, their size ranges, and enterprise considerations (ranges are estimates and may vary).

Context window implications

The context window sets a hard limit on the number of input and output tokens per call, including any tokens used by reasoning models for chain-of-thought reasoning. If the total exceeds this limit, the response is truncated, or the request fails outright.

Figure 2: Illustration of context window limitations leading to output truncation in a multi-turn conversation.⁵

For applications that maintain long conversations, every additional turn pushes more history into the input. Without intervention, input tokens grow linearly with conversation length, and so does the bill. API users typically address this in one of three ways:

• Prompt caching. OpenAI, Anthropic, Google, and DeepSeek all cache repeated prompt prefixes server-side and bill cache hits at a fraction of the standard input rate, typically 10 to 50 percent of the cache-miss price. For applications that reuse a long system prompt or conversation prefix, caching can cut input cost by an order of magnitude.
• Rolling window or RAG. Drop the oldest turns once a threshold is hit, or retrieve only relevant past messages from a vector store on each call.
• Summarization. Periodically condense older turns into a summary instead of resending them verbatim.

For agentic workloads such as coding sessions or deep research, modern coding agents handle this automatically in session. Claude Code, for example, ships with context compaction: when the conversation approaches the limit, it summarizes older messages into a condensed version while keeping recent turns intact. Subsequent turns send only the summary plus recent context back to the model.

The pricing impact is direct. On per-token APIs, prompt caching and compaction cap how large each call’s input grows, so cost-per-turn stays predictable across long sessions. On flat-rate subscriptions like Claude Pro, ChatGPT Plus, or Kimi Moderato, compaction stretches daily and weekly usage limits because each call carries less context. A coding session that would otherwise burn through a 5-hour rate limit can run longer when older turns get compressed.

The trade-off is that any form of summarization is lossy. The summary may drop details that turn out to matter later, forcing the user to re-supply them

Max output tokens

Max output tokens caps the length of a model’s response. While many documentations mention that it can be adjusted using the max_tokens parameter, it is crucial to review the documentation of the specific API being used to identify the correct parameter. It should be adjusted according to the specific needs:

If set too low, it may result in incomplete outputs, causing the model to cut off responses before delivering the full answer.

If set too high, depending on the temperature (a parameter that controls response creativity), it can lead to unnecessarily verbose outputs, longer response times, and increased cost.

Therefore, it is a parameter that requires careful consideration to optimize resource usage while balancing output quality, cost, and performance.

Table 3: Example input prompts and estimated token counts per content type.

*This assumes that each model produces responses with an equal number of output tokens, although the token count for both input and output may vary depending on each model’s tokenization; the number has been kept constant here for each model.

The LLM API price calculator can be used to determine the total cost per model when generating content types from Table 2 via the API, using the sample prompts provided in Table 3. Additionally, it can be used to calculate costs for custom cases beyond the suggested content types.

LLM API price calculator

You can calculate your total cost by filling out these 3 values below and sorting the results by input cost, output cost, total cost, or alphabetically in increasing or decreasing order:

Note: The default ranking is based on the total cost.

Comparing LLM subscription plans

Non-technical users may prefer to use the UI rather than the API. In 2026, most provider subscriptions bundle far more than a chat interface. Coding agents like Claude Code, Codex, Kimi Code, and Mistral Vibe ship inside Pro-tier plans. For developers and heavy users, the right $10–$200 subscription often replaces what would otherwise be a separate coding IDE subscription, a per-token API budget, and a video or research tool combined.

OpenAI

Free plan includes access to GPT-5.5 instant with capped daily usage, standard voice mode, limited uploads, and basic image generation. Contextual ads now appear in a few regions, including the U.S.

• ChatGPT Go ($8/month) is a low-cost, ad-supported plan that offers roughly 10x the free tier’s messages, file uploads, image creation, and full access to GPT-5.5.
• ChatGPT Plus ($20/month) includes extended usage limits, access to GPT-5.5 and current reasoning models, advanced voice mode, Codex agent, image and video generation, and early-access features.

Pro plan has two tiers as of April 2026:

• ChatGPT Pro ($100/month) provides the same model lineup as the $200 tier (including GPT-5.5 Pro and the latest reasoning models) at roughly 5x Plus usage limits. Bundled apps: Codex with 5x Plus usage, more Deep Research runs, and full Sora access.
• ChatGPT Pro ($200/month) provides the highest individual usage limits (about 20x Plus), 250 Deep Research runs per month, advanced voice with video and screensharing, Codex with maximum usage boost, Sora, and Operator preview (U.S. only).

Both Pro tiers include priority access during peak hours. Codex pricing on Plus, Pro, and Business shifted from per-message to API-token-aligned usage in April 2026.

• Business plan ($20/user/month annual or $25/user/month) is OpenAI’s plan for small and mid-sized teams (formerly ChatGPT Team, renamed in August 2025). It adds higher message limits, admin console, SSO, training-excluded team data, and shared credit pools for advanced features. Bundled apps: Codex with shared workspace credits and the option to assign separate Codex-only seats at flexible, usage-based pricing. Minimum of 2 seats.

• The Enterprise plan (custom pricing) provides high-speed model access, expanded context windows, enterprise-grade data controls, domain verification, analytics, and audit logs. Bundled apps: Codex with shared credit pool, optional Codex-only seats, and Operator access.

Anthropic (Claude)

Free plan includes web and mobile access, basic analysis, access to Claude Sonnet 4.6, and document uploading. Daily usage is capped, and Opus models are not available.

• Pro plan ($20/month, or $17/month billed annually) provides access to all Claude models, including Opus 4.7 and Sonnet 4.6, roughly 5x more usage than Free, project organization, and priority access during peak hours. Bundled apps: Claude Code (Anthropic’s coding agent in the terminal and IDE) and Cowork (Research mode), both sharing the same usage pool as the chat. As of May 2026, Claude Code’s five-hour rate limits doubled, and the peak-hour reduction was removed.

• Max 5x plan ($100/month) provides about 5x more usage than Pro, priority access to the newest features and models, and full Claude Code access at the higher Max usage tier.

• Max 20x plan ($200/month) provides about 20x more usage than Pro, maximum priority access, and full Claude Code access. Designed for daily power users running Claude Code workloads.

Team plan offers two seat types and supports 5–150 members:

• Standard seat: $20/user/month annual ($25/user/month monthly). Includes base features, standard usage limits, and Claude Code access.
• Premium seat: $100/user/month annual ($125/user/month monthly). Everything in Standard, plus higher usage limits for power users running heavier Claude Code workloads.

Bundled apps: Claude Code and Cowork are included with every Team seat (Standard and Premium); the difference lies in the usage allowance, not access. Both seat types include central billing, collaboration tools, and admin controls.

• Enterprise plan (custom pricing) provides expanded context windows, SSO, domain capture, role-based access, SCIM, audit logs, and data integrations. Bundled apps: on new and self-serve Enterprise plans, Claude Code and Cowork are included with every seat; older Enterprise contracts may distinguish between Chat-only seats and Chat + Claude Code seats with usage-based billing.

Google (Gemini)

The free plan provides access to Gemini 3 Flash and varying access to Gemini 3.1 Pro, basic image generation, Deep Research, Gemini Live, Canvas, and Gems. Bundled apps: NotebookLM (research and writing assistant) and Flow (limited Veo 3.1 access for AI filmmaking).

Google uses regional pricing, so pricing can vary by region.

• Google AI Plus ($7.99/month, U.S.) is the entry paid tier. Bundled apps: enhanced Gemini 3.1 Pro access in the chat, image generation with Nano Banana Pro, Veo 3.1 Lite video generation, Flow with limited Veo 3.1, NotebookLM with more Audio Overviews, Gemini in Gmail, Docs and Vids, and early-access Gemini in Chrome. Includes 200 GB of storage.

• Google AI Pro ($19.99/month, U.S.) provides higher usage limits for Gemini 3.1 Pro and 5 TB of storage. Bundled apps: Jules (asynchronous coding agent), Gemini Code Assist and Gemini CLI for IDEs, Google Antigravity (agentic development platform), NotebookLM with 5x Audio Overviews, Deep Research, Veo 3.1 Lite video, and Google Home Premium (Standard plan).

• Google AI Ultra ($249.99/month, with a U.S. introductory offer of $124.99/month for the first three months) provides the highest usage limits across all features and 30 TB of storage. Bundled apps: full Veo 3.1 video generation, Deep Think reasoning, Gemini Agent (U.S. only), Project Mariner agentic browsing, Project Genie (interactive world model), Jules at 20x Pro limits, highest-tier Antigravity, NotebookLM at maximum capability, Google Home Premium (Advanced plan), and a YouTube Premium individual subscription.

Microsoft Copilot

The free plan (Copilot Chat) is available at no additional cost for all Microsoft Entra users with an eligible Microsoft 365 subscription. It includes basic Copilot chat across Microsoft apps without the deeper in-document features.

• Copilot Pro ($20/month) adds priority model access, image-generation boosts, and full Copilot integration with Word, Excel, PowerPoint, Outlook, and OneNote, plus Copilot in Designer for image and document layouts. It requires an active Microsoft 365 Personal or Family subscription. Microsoft has also folded most Pro features into a new Microsoft 365 Premium plan ($19.99/month) that bundles Office apps, 1 TB of OneDrive, and Copilot into a single subscription.

• Microsoft 365 Copilot Business ($18/user/month promotional rate through June 30, 2026, then $21/user/month annual; $25.20/user/month monthly) adds Copilot across Microsoft 365 apps, Teams integration, and admin controls. Bundled apps: Copilot Studio Lite for building lightweight agents, Copilot in SharePoint, and Copilot Pages for collaborative drafts. Limited to organizations with up to 300 users.

• Microsoft 365 Copilot Enterprise ($30/user/month, annual commitment) provides advanced security, compliance, and analytics on top of Business features. Bundled apps: full Copilot Studio for custom agent development, Copilot in Microsoft Purview and Intune for IT and security workflows, and enterprise-grade governance over deployed agents.

xAI (Grok)

The free plan provides limited Grok access with approximately 10 requests every two hours.

• SuperGrok Lite ($10/month) is the entry paid tier. It includes 2x longer conversations, increased rate limits, and AI image and video creation. Bundled apps: 1 AI agent on Expert mode and Grok Imagine for image and video generation.

• SuperGrok ($30/month, or $300/year) includes enhanced reasoning, lightning-fast replies, longer file uploads, and the staged rollout of Grok 4.3. Bundled apps: 4 AI agents on Expert mode running in parallel, DeepSearch for live web research, Big Brain mode for extended thinking, Voice mode for spoken chat, and 20x more Grok Imagine image and video generations including HD 720p 30-second video.

• SuperGrok Heavy ($300/month) provides full access to Grok 4.3, Grok 4 Heavy (multi-agent reasoning with a 256K context window), maximum rate limits, priority access during peak load, and early previews of upcoming xAI features. Bundled apps: maximum agent concurrency on Expert mode, full DeepSearch, Big Brain, Voice, and Grok Imagine quotas.

Grok is also bundled into X subscriptions: X Premium ($8/month) is the cheapest paid path to Grok inside the X app and includes verified status and ad-free browsing. X Premium+ ($40/month) bundles Grok with full creator monetization, the staged Grok 4.3 rollout, and the same Grok agent and DeepSearch capabilities at the X Premium+ usage tier.

Moonshot AI (Kimi)

Kimi’s consumer plans are named after musical tempo markings, from slowest to fastest. International pricing is in USD; Chinese users pay in CNY at lower rates.

• Adagio (Free) provides unlimited basic conversations with 6 agent uses, capped Deep Research queries, and basic OK Computer agent tasks.

• Moderato ($19/month) adds Kimi K2.6 in chat and agent tasks plus expanded Deep Research sessions. Bundled apps: Kimi Code (terminal-first AI coding agent with 300–1,200 API calls per 5-hour window) at 1x credit, plus Slides and Websites authoring tools.

• Allegretto ($39/month) provides higher usage on everything in Moderato. Bundled apps: Agent Swarm (parallel subagent orchestration with 100 sub-agents and ~1,500 coordinated steps in K2.5, scaling to 300 sub-agents and 4,000 steps in K2.6), Kimi Claw cloud deployment for heterogeneous agent groups with persistent memory, and 5x Kimi Code credits.

• Allegro ($99/month) provides Agent Swarm with 120 monthly uses, 15x Kimi Code credits, and 12,000 Pro Data requests for research-heavy workflows.

• Vivace ($199/month) provides Agent Swarm with 240 monthly uses and up to 8 parallel subagents, 30x Kimi Code credits, and 24,000 Pro Data requests. Targeted at heavy research and agentic workloads.

Membership does not include API usage, which is billed separately per token.

MiniMax

MiniMax separates its consumer Agent product from its coding-focused subscriptions, both of which sit on top of the underlying M2.x model family.

MiniMax Agent plans (autonomous multi-step research, programming, and Office workflows):

• Free: 1,000 starter credits valid for 3 days, plus 200 daily credits that refresh and roll over.
• Basic ($39/month): 5,000 credits per month (~30 Pro-mode tasks), peak-hour priority, watermark removal, custom domain, 1 MaxClaw, and 1 MaxHermes 24/7 cloud deployments.
• Pro ($119/month): 20,000 credits per month (~120 Pro-mode tasks), 3 MaxClaw and 1 MaxHermes deployments, plus all Basic perks.
• Ultra ($219/month): 40,000 credits per month (~240 Pro-mode tasks), the same deployment count as Pro, and the highest priority.
• Team (custom): central billing and admin controls for organizations.

MiniMax Coding Plan (separate, layered on top of the API for developers; powered by MiniMax M2.x):

• $10/month: 100 prompts per 5-hour window.
• $20/month (Plus): 300 prompts per 5-hour window.
• $50/month (Max): 1,000 prompts per 5-hour window.

The Coding Plan ships with predictable prompt quotas rather than token-based billing, making it one of the cheapest paths to a frontier coding model when paired with a CLI like Cline or Kilo Code.

Mistral AI

Free plan (Le Chat) includes web browsing, basic file analysis, image generation, fast Flash responses, group chats organized into projects, up to 500 saved memories, and 40+ enterprise connectors.

• Pro plan ($14.99/user/month) includes more messages and web searches, more extended thinking and Deep Research reports, 15 GB of document storage, up to 1,000 projects, and state-of-the-art image generation. Bundled apps: Mistral Vibe (Mistral’s coding agent for all-day development, with pay-as-you-go beyond included quota). Mistral also offers a Student tier at $7.04/user/month with the same Pro features.

• Team plan ($24.99/user/month) includes everything in Pro with up to 30 GB of storage per user, central billing, role-based access control, domain name verification, and data export. Bundled apps: Mistral Vibe at the team usage tier with shared admin controls.

• Enterprise plan (custom pricing) provides secure deployment options, including self-hosted and private cloud, SAML SSO, audit logs, premium support, and detailed analytics. Bundled apps: Mistral Vibe with on-premise deployment options for regulated workloads.

DeepSeek

DeepSeek does not offer traditional subscription plans. Web and mobile chat access to the latest models (currently DeepSeek V4-Flash and V4-Pro) is free for all users, with fair-use throttling that resets daily.

API access is pay-per-token only. V4-Flash is priced at $0.14 per million input tokens (cache miss) and $0.28 per million output tokens, with cache hits served at roughly 1/50th of the input rate.

Meta (Muse Spark)

Meta does not currently sell a consumer subscription for its AI assistant. Muse Spark, the first model from Meta Superintelligence Labs (launched April 8, 2026), is a natively multimodal reasoning model with tool use, visual chain-of-thought, and multi-agent orchestration. It powers Meta AI inside WhatsApp, Instagram, Facebook, Messenger, the Meta AI app, and Ray-Ban Meta glasses, all at no cost to end users.

API access is currently in private preview for select developers and enterprises, with no published pricing. Meta has indicated that broader availability and pricing will follow.

Using multiple language models

A tool like OpenRouter allows the same prompt to be sent to multiple models simultaneously. The responses, token consumption, response time, and pricing can then be compared to determine which model is most suitable for the task.

Figure 3: Interface showcasing a prompt sent to multiple Large Language Models (LLMs), including R1, Mistral Small 3, GPT-4o-mini, and Claude 3.5 Sonnet.⁶

Benefits and challenges

• Increased adaptability and efficiency: Orchestration enhances responsiveness, enabling real-time assessment of model efficiency and identifying a cost-effective model and potential savings.
• Prompt sensitivity and optimization: Identical prompts can elicit vastly different outputs across models, necessitating prompt engineering tailored to each model to achieve desired results, adding to development and maintenance complexity.

Pricing mechanics & hidden costs

Reasoning tokens vs. output tokens

A growing number of providers have introduced reasoning models that spend additional compute to perform chain-of-thought reasoning internally. These models may use a separate “reasoning token” class (distinct from standard output tokens), which typically incurs significantly higher costs.

For example, models like o3, o4-mini, or Claude Sonnet 4.6 with extended thinking generate internal reasoning traces even when you do not explicitly request them. These internal tokens count toward your bill and can substantially increase cost, especially in long analytical tasks such as legal review, data analysis, or multi-step reasoning.

This makes it essential to:

• Choose a reasoning model only when accuracy substantially outweighs cost.
• Disable the chain-of-thought or set a shorter max output token count when possible.
• Test the same task on non-reasoning models to see if performance is comparable at a fraction of the price.

Since reasoning models can generate 10-30x more thinking tokens per request, it is critical to understand this distinction for cost planning.

Architecture-driven pricing differences

LLM architectures directly influence model efficiency and, therefore, API pricing. For example:

• Mixture-of-Experts (MoE) models activate only a subset of parameters per request, reducing compute cost and allowing providers to offer lower per-token rates.
• Speculative decoding pairs a smaller draft model with a larger one, improving throughput and lowering cost for deterministic tasks.
• Quantized variants (e.g., 4-bit or 8-bit) can perform inference at lower precision, enabling lower pricing for locally deployed or cloud-hosted versions.

Understanding these architectural choices helps users predict not only pricing differences but also latency, quality, and how a model scales under production workloads.

Operational costs beyond API fees

While per-token pricing is the primary cost driver, many production deployments incur additional costs beyond API usage:

• Embeddings and vector databases: Storing and retrieving vectors (e.g., Pinecone, Weaviate, ChromaDB) adds cost per query and per GB of storage.
• Reranking and post-processing models: Many applications use smaller models for summarization, filtering, or classification before sending a final request to a bigger model.
• Caching layers: Providers like OpenAI now offer prompt-level caching, but local caching infrastructure may require additional compute.
• Logging, monitoring, and auditing: Enterprises often incur costs for token-level monitoring, latency tracking, and security audits.

These hidden costs often account for 20–40% of total LLM operational expenses and should be considered when evaluating pricing structures.

Enterprise-specific pricing considerations

Many LLM vendors charge additional fees for enterprise-grade security and compliance features, such as:

• Single-tenant deployments
• Dedicated GPU clusters
• Enhanced SLAs (e.g., uptime, latency guarantees)
• Data residency and regional controls
• SOC2, HIPAA, or GDPR compliance modes

These offerings can increase costs significantly but are essential for regulated industries such as healthcare, finance, legal services, and public institutions.

Future trends in LLM pricing

Commoditization of general models

General-purpose language models have become less expensive as competition has intensified and open-source options have expanded. Capabilities such as summarization, fundamental question answering, and standard content generation require less specialized computation, which has led providers to aggressively lower per-token rates through 2025 and into 2026.

• Growing availability of efficient open-source models (Llama 4, Qwen3, DeepSeek V4).
• Multiple price cuts from major providers. OpenAI reduced o3 pricing from $10 to $2 per million input tokens. Google cut Gemini 2.5 Pro input pricing by 50%.
• More generous context windows as a standard feature rather than a differentiator. Multiple models now offer 1M+ token windows.

This stage mirrors the early cloud compute market, where basic capacity became affordable as providers scaled.

Premium pricing for reasoning and multimodal models

Advanced reasoning and multimodal systems continue to command a premium. Models like o3-pro ($20/$80 per million tokens) and o1-pro ($150/$600 per million tokens) cost 10-100x more than standard models.

• Higher compute requirements for complex reasoning chains.
• Demand for accuracy-sensitive workflows in legal, medical, and financial domains.
• Clear distinction between commodity-language tasks and high-precision tasks.

This two-tier market is now established: inexpensive general models for routine work and premium models for tasks that depend on stronger reasoning performance.

Growth of subscription-based pricing

The market has shifted toward subscription plans that bundle AI access with existing products. Google bundles Gemini with Google One storage. Microsoft bundles Copilot with Microsoft 365. xAI bundles Grok with X Premium. OpenAI introduced ChatGPT Go at $8/month with ad subsidies.

These bundled approaches make AI access a feature of an existing subscription rather than a standalone purchase.

Expansion of SLA-based pricing tiers

Enterprises with strict reliability or regulatory requirements are adopting service levels similar to those used in cloud infrastructure. These tiers differentiate on uptime guarantees, latency expectations, data residency options, and support response times.

• Clear structure for organizations with varied operational needs.
• Standard, business, and mission-critical tiers.
• Pricing aligned with performance expectations.

FAQs

Reference Links

1.

OpenAI Platform

2.

[1508.07909] Neural Machine Translation of Rare Words with Subword Units

3.

[1810.04805] BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

4.

[1808.06226] SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing

5.

OpenAI Platform

6.

OpenRouter

Cem Dilmegani

Principal Analyst

Follow On

Cem has been the principal analyst at AIMultiple since 2017. AIMultiple informs hundreds of thousands of businesses (as per similarWeb) including 55% of Fortune 500 every month.

Cem's work has been cited by leading global publications including Business Insider, Forbes, Washington Post, global firms like Deloitte, HPE and NGOs like World Economic Forum and supranational organizations like European Commission. You can see more reputable companies and resources that referenced AIMultiple.

Throughout his career, Cem served as a tech consultant, tech buyer and tech entrepreneur. He advised enterprises on their technology decisions at McKinsey & Company and Altman Solon for more than a decade. He also published a McKinsey report on digitalization.

He led technology strategy and procurement of a telco while reporting to the CEO. He has also led commercial growth of deep tech company Hypatos that reached a 7 digit annual recurring revenue and a 9 digit valuation from 0 within 2 years. Cem's work in Hypatos was covered by leading technology publications like TechCrunch and Business Insider.

Cem regularly speaks at international technology conferences. He graduated from Bogazici University as a computer engineer and holds an MBA from Columbia Business School.

View Full Profile

Be the first to comment

Your email address will not be published. All fields are required.

Name

Email Address

Comment

0/450

Post Comment