LLM API pricing: OpenAI, Claude and Gemini per 1M tokens

Standard pay-as-you-go prices in USD per 1 million tokens, excluding tax, for 12 models we track. Where a model has prompt-length tiers, each tier is listed with its condition. Batch, priority and regional-endpoint variants are not shown here. Products billed per request are listed after the token tables.

This is not a full market list: we only include models whose prices we have checked on the vendor's own pricing page. Introductory prices are flagged with their end date.

OpenAI API pricing

OpenAI API prices run from $0.10 input / $0.50 output per 1M tokens (GPT-6 Luna) to $10/$50 (GPT-6 Astra) for prompts up to 272K tokens. In between, GPT-6.1 Sol is $2/$10. Longer prompts cost more. Batch is 50% off.

Estimate your monthly OpenAI API bill →

OpenAI prices per 1M tokens (USD, excl. tax)
Model	Input / 1M	Cached input / 1M	Output / 1M	Context window
GPT-6 Astra gpt-6-astra Monthly cost →	$10 ≤272K input tokens (short context) $20 >272K input tokens (long context)	$1 ≤272K input tokens (short context) $2 >272K input tokens (long context)	$50 ≤272K input tokens (short context) $75 >272K input tokens (long context)	not published
GPT-6.1 Sol gpt-6.1-sol Monthly cost →	$2 ≤272K input tokens (short context) $4 >272K input tokens (long context)	$0.10 ≤272K input tokens (short context) $0.20 >272K input tokens (long context)	$10 ≤272K input tokens (short context) $15 >272K input tokens (long context)	not published
GPT-6 Luna gpt-6-luna Monthly cost →	$0.10 ≤272K input tokens (short context) $0.20 >272K input tokens (long context)	$0.01 ≤272K input tokens (short context) $0.02 >272K input tokens (long context)	$0.50 ≤272K input tokens (short context) $0.75 >272K input tokens (long context)	not published

Last checked Oct 19, 2026Source: developers.openai.com ↗

Anthropic API pricing

Claude API prices start at $0.10 input / $0.50 output per 1M tokens (Claude Haiku 5.5, prompts up to 100K tokens) and go up to $10/$50 (Claude Fable 5.1). In between, Claude Sonnet 5.5 is $2/$10 and Claude Opus 5.5 is $4/$20. Longer prompts cost more. Batch is 50% off.

Estimate your monthly Anthropic API bill →

Anthropic prices per 1M tokens (USD, excl. tax)
Model	Input / 1M	Cached input / 1M	Output / 1M	Context window
Claude Fable 5.1 claude-fable-5.1 Monthly cost →	$10	$0.25	$50	1,000,000
Claude Opus 5.5 claude-opus-5.5 Monthly cost →	$4	$0.20	$20	1,000,000
Claude Sonnet 5.5 claude-sonnet-5.5 Monthly cost →	$2	$0.10	$10	1,000,000
Claude Haiku 5.5 claude-haiku-5.5 Monthly cost →	$0.10 prompt ≤100,000 tokens $0.50 prompt >100,000 tokens	$0.01 cache hits, prompt ≤100,000 tokens $0.05 cache hits, prompt >100,000 tokens	$0.50 prompt ≤100,000 tokens $2.50 prompt >100,000 tokens	not published

Last checked Oct 19, 2026Source: platform.claude.com ↗

Google API pricing

Gemini API paid-tier prices start at $0.30 input / $2.50 output per 1M tokens (Gemini 3.5 Flash-Lite). Some models have dated rates: Gemini 4 Argon is $2/$10 until Nov 30, 2026, with higher rates scheduled from Dec 1, 2026, and Gemini 3.8 Flash is $0.75/$3.75 until Dec 31, 2026, with higher rates scheduled from Jan 1, 2027. Batch is 50% off.

Estimate your monthly Google Gemini API bill →

Google prices per 1M tokens (USD, excl. tax)
Model	Input / 1M	Cached input / 1M	Output / 1M	Context window
Gemini 4 Argon gemini-4-argon Monthly cost → Announced from Dec 1, 2026: input $4.00, output $20.00	$2Introductory price until Nov 30, 2026	$0.10Introductory price until Nov 30, 2026	$10Introductory price until Nov 30, 2026	not published
Gemini 3.8 Flash gemini-3.8-flash Monthly cost → Announced from Jan 1, 2027: input $1.50, output $7.50	$0.75Introductory price until Dec 31, 2026	$0.08Introductory price until Dec 31, 2026	$3.75Introductory price until Dec 31, 2026	not published
Gemini 3.5 Flash-Lite gemini-3.5-flash-lite Monthly cost →	$0.30	$0.03	$2.50	not published

Last checked Oct 19, 2026Source: ai.google.dev ↗

xAI API pricing

xAI Grok API prices run from $1.25 input / $2.50 output per 1M tokens (Grok 4.3) to $2/$6 (Grok 4.7) for prompts under 200K tokens. Prompts of 200K tokens or more are billed at a higher rate for all tokens in the request.

Estimate your monthly xAI API bill →

xAI prices per 1M tokens (USD, excl. tax)
Model	Input / 1M	Cached input / 1M	Output / 1M	Context window
Grok 4.7 grok-4.7 Monthly cost →	$2 prompt < 200k tokens $4 prompt ≥ 200k tokens (all tokens in the request billed at this rate)	$0.50 prompt < 200k tokens $1 prompt ≥ 200k tokens (all tokens in the request billed at this rate)	$6 prompt < 200k tokens $12 prompt ≥ 200k tokens (all tokens in the request billed at this rate)	500,000
Grok 4.3 grok-4.3 Monthly cost →	$1.25 prompt < 200k tokens $2.50 prompt ≥ 200k tokens (all tokens in the request billed at this rate)	$0.20 prompt < 200k tokens $0.40 prompt ≥ 200k tokens (all tokens in the request billed at this rate)	$2.50 prompt < 200k tokens $5 prompt ≥ 200k tokens (all tokens in the request billed at this rate)	1,000,000

Last checked Oct 19, 2026Source: docs.x.ai ↗

Perplexity API pricing

Perplexity’s Search API costs $5 per 1,000 requests, or $1 per 1,000 with Fast Search, with no token costs. Support for Sonar Chat Completions ended on Sep 27, 2026; Perplexity now handles those requests through its Agent API.

Estimate your monthly Perplexity API bill →

These products are billed per request or per tool call, not per token. Sonar Pro and Sonar were retired on Sep 27, 2026 and are no longer listed.

Perplexity per-request and per-call prices (USD, excl. tax)
Product	Price	What it covers
Agent API tool fees	$0 per call	web_search (standard), per invocation
Agent API tool fees	$0 per call	web_search (Fast Search), per invocation
Agent API tool fees	$0 per call	fetch_url, per invocation
Agent API tool fees	$0.03 per call	sandbox, per session (up to 20 min active use)
Search API	$5 per 1,000 requests	Search API, per 1K successful requests (no token costs)
Search API	$1 per 1,000 requests	Search API with Fast Search (search_type: 'fast'), per 1K requests

Last checked Oct 19, 2026Source: docs.perplexity.ai ↗

API or subscription?

Paying $20/mo for ChatGPT Plus or Claude Pro? See when the API is cheaper →

Estimate your costs

Paste your prompt into the token counter to see what it would cost with each of these models. For a whole workload (requests per day, caching, batch), use the LLM API cost calculator.