Verified 17 October 2026

LLM API pricing, per million tokens.

List prices for 95 models from 7 providers, checked against each provider's own pricing page. Input, cached input, cache write, output and batch, in one place. This is also the default rate card Agent Ping prices every agent run with, so what you see here is what lands on your Spend page.

Anthropic

Source: platform.claude.com →

Cache write is the 5-minute cache; the 1-hour cache is 2x input. Cache reads are 0.1x input on most models, 0.05x on Opus 5.5 and Sonnet 5.5, 0.025x on Fable 5.1 and Mythos 5.1. Batch API is 50% off. Haiku 5.5 prompts over 100k tokens bill at $0.50 / $2.50.

Model Input Cached input Cache write Output Batch Notes
Claude Fable 5.1claude-fable-5-1 $10 $0.25 $12.5 $50 50% off
Claude Mythos 5.1claude-mythos-5-1 $10 $0.25 $12.5 $50 50% off Limited availability.
Claude Fable 5claude-fable-5 $10 $1 $12.5 $50 50% off
Claude Mythos 5claude-mythos-5 $10 $1 $12.5 $50 50% off Limited availability.
Claude Opus 5.5claude-opus-5-5 $4 $0.2 $5 $20 50% off
Claude Opus 5claude-opus-5 $5 $0.5 $6.25 $25 50% off
Claude Opus 4.8claude-opus-4-8 $5 $0.5 $6.25 $25 50% off
Claude Opus 4.7claude-opus-4-7 $5 $0.5 $6.25 $25 50% off
Claude Opus 4.6claude-opus-4-6 $5 $0.5 $6.25 $25 50% off
Claude Opus 4.5claude-opus-4-5 $5 $0.5 $6.25 $25 50% off
Claude Opus 4.1retiredclaude-opus-4-1 $15 $1.5 $18.75 $75 50% off Retired on the Claude API; still on Bedrock and Google Cloud.
Claude Opus 4retiredclaude-opus-4 $15 $1.5 $18.75 $75 50% off Retired on the Claude API; still on Google Cloud.
Claude Sonnet 5.5claude-sonnet-5-5 $2 $0.1 $2.5 $10 50% off
Claude Sonnet 5claude-sonnet-5 $2 $0.2 $2.5 $10 50% off
Claude Sonnet 4.6claude-sonnet-4-6 $3 $0.3 $3.75 $15 50% off
Claude Sonnet 4.5legacyclaude-sonnet-4-5 $3 $0.3 $3.75 $15 50% off
Claude Sonnet 4retiredclaude-sonnet-4 $3 $0.3 $3.75 $15 50% off Retired on the Claude API; still on Bedrock and Google Cloud.
Claude Haiku 5.5claude-haiku-5-5 $0.1 $0.01 $0.125 $0.5 50% off Prompts over 100k tokens bill at $0.50 input / $2.50 output.
Claude Haiku 4.5legacyclaude-haiku-4-5 $1 $0.1 $1.25 $5 50% off
Claude Haiku 3.5retiredclaude-3-5-haiku $0.8 $0.08 $1 $4 50% off Retired on the Claude API; still on Google Cloud.

USD per 1,000,000 tokens, standard tier. A dash under cached input means the provider publishes no cached rate; under cache write, no per-token write charge.

OpenAI

Source: developers.openai.com →

No charge for cache writes. Batch API is 50% off. GPT-5.4 and later, GPT-5.5 and the GPT-6 family bill prompts over 272k input tokens at roughly 2x input and 1.5x output. Pro models have no cached-input rate.

Model Input Cached input Cache write Output Batch Notes
GPT-6 Astragpt-6-astra $10 $1 — $50 50% off Over 272k input tokens: $20 / $75.
GPT-6.1 Solgpt-6.1-sol $2 $0.1 — $10 50% off Over 272k input tokens: $4 / $15.
GPT-6 Solgpt-6-sol $2 $0.2 — $10 50% off Over 272k input tokens: $4 / $15.
GPT-6 Lunagpt-6-luna $0.1 $0.01 — $0.5 50% off Over 272k input tokens: $0.20 / $0.75.
GPT-5.6 Solgpt-5.6-sol $4 $0.4 — $20 50% off Over 272k input tokens: $8 / $30.
GPT-5.6 Terragpt-5.6-terra $2 $0.2 — $12 50% off Over 272k input tokens: $4 / $18.
GPT-5.6 Lunagpt-5.6-luna $0.2 $0.02 — $1.2 50% off Over 272k input tokens: $0.40 / $1.80.
GPT-5.5gpt-5.5 $5 $0.5 — $30 50% off Over 272k input tokens: $10 / $45.
GPT-5.5 Progpt-5.5-pro $30 — — $180 50% off No cached-input rate. Over 272k input tokens: $60 / $270.
GPT-5.4gpt-5.4 $2.5 $0.25 — $15 50% off Over 272k input tokens: $5 / $22.50.
GPT-5.4 Minigpt-5.4-mini $0.75 $0.075 — $4.5 50% off
GPT-5.4 Nanogpt-5.4-nano $0.2 $0.02 — $1.25 50% off
GPT-5.4 Progpt-5.4-pro $30 — — $180 50% off No cached-input rate. Over 272k input tokens: $60 / $270.
GPT-5.2gpt-5.2 $1.75 $0.175 — $14 50% off
GPT-5.2 Progpt-5.2-pro $21 — — $168 50% off No cached-input rate.
GPT-5.1gpt-5.1 $1.25 $0.125 — $10 50% off
GPT-5gpt-5 $1.25 $0.125 — $10 50% off
GPT-5 Minigpt-5-mini $0.25 $0.025 — $2 50% off
GPT-5 Nanogpt-5-nano $0.05 $0.005 — $0.4 50% off
GPT-5 Progpt-5-pro $15 — — $120 50% off No cached-input rate.
GPT-4.1legacygpt-4.1 $2 $0.5 — $8 50% off
GPT-4.1 Minilegacygpt-4.1-mini $0.4 $0.1 — $1.6 50% off
GPT-4.1 Nanolegacygpt-4.1-nano $0.1 $0.025 — $0.4 50% off
GPT-4olegacygpt-4o $2.5 $1.25 — $10 50% off
GPT-4o Minilegacygpt-4o-mini $0.15 $0.075 — $0.6 50% off
o3legacyo3 $2 $0.5 — $8 50% off
o3-prolegacyo3-pro $20 — — $80 50% off No cached-input rate.
o3-minilegacyo3-mini $1.1 $0.55 — $4.4 50% off
o4-minilegacyo4-mini $1.1 $0.275 — $4.4 50% off
o1legacyo1 $15 $7.5 — $60 50% off

USD per 1,000,000 tokens, standard tier. A dash under cached input means the provider publishes no cached rate; under cache write, no per-token write charge.

Google Gemini

Source: ai.google.dev →

Paid tier, text input. Context caching bills storage per hour instead of a per-token write. Batch is 50% off. Pro models bill prompts over 200k tokens at 2x input and 1.5x output. Gemini 4 Argon list price doubles on 1 December 2026; the 3.6 and 3.8 Flash introductory prices end 1 January 2027.

Model Input Cached input Cache write Output Batch Notes
Gemini 4 Argongemini-4-argon $2 $0.1 — $10 50% off Rises to $4 / $20 on 1 December 2026. Not on the free tier.
Gemini 3.8 Flashgemini-3.8-flash $0.75 $0.075 — $3.75 50% off Introductory; $1.50 / $7.50 from 1 January 2027.
Gemini 3.6 Flashlegacygemini-3.6-flash $0.75 $0.075 — $3.75 50% off Introductory; $1.50 / $7.50 from 1 January 2027.
Gemini 3.5 Flash-Litegemini-3.5-flash-lite $0.3 $0.03 — $2.5 50% off
Gemini 3.1 Flash-Litegemini-3.1-flash-lite $0.25 $0.025 — $1.5 50% off
Gemini 3.1 Pro Previewgemini-3.1-pro-preview $2 $0.2 — $12 50% off Over 200k tokens: $4 / $18.
Gemini 3 Flash Previewlegacygemini-3-flash-preview $0.5 $0.05 — $3 50% off
Gemini 2.5 Prolegacygemini-2.5-pro $1.25 $0.125 — $10 50% off Over 200k tokens: $2.50 / $15.
Gemini 2.5 Flashlegacygemini-2.5-flash $0.3 $0.03 — $2.5 50% off
Gemini 2.5 Flash-Litelegacygemini-2.5-flash-lite $0.1 $0.01 — $0.4 50% off

USD per 1,000,000 tokens, standard tier. A dash under cached input means the provider publishes no cached rate; under cache write, no per-token write charge.

xAI

Source: docs.x.ai →

Every model bills a prompt of 200k tokens or more at 2x on all tokens in the request. Priority processing is also 2x. No batch tier.

Model Input Cached input Cache write Output Batch Notes
Grok 4.7grok-4.7 $2 $0.5 — $6 — Prompts of 200k tokens or more: $4 / $12.
Grok 4.6grok-4.6 $2 $0.5 — $6 — Prompts of 200k tokens or more: $4 / $12.
Grok 4.5legacygrok-4.5 $2 $0.3 — $6 — Prompts of 200k tokens or more: $4 / $12.
Grok 4.3grok-4.3 $1.25 $0.2 — $2.5 — Prompts of 200k tokens or more: $2.50 / $5.
Grok 4.20 Reasoninglegacygrok-4.20-0309-reasoning $1.25 $0.2 — $2.5 — Prompts of 200k tokens or more: $2.50 / $5.
Grok 4.20 Non-reasoninglegacygrok-4.20-0309-non-reasoning $1.25 $0.2 — $2.5 — Prompts of 200k tokens or more: $2.50 / $5.
Grok 4.20 Multi-agentlegacygrok-4.20-multi-agent-0309 $1.25 $0.2 — $2.5 — Prompts of 200k tokens or more: $2.50 / $5.
Grok Build 0.1grok-build-0.1 $1 $0.2 — $2 — Prompts of 200k tokens or more: $2 / $4.

USD per 1,000,000 tokens, standard tier. A dash under cached input means the provider publishes no cached rate; under cache write, no per-token write charge.

Mistral AI

Source: docs.mistral.ai →

Cached input is 0.1x. Batch is 50% off. Mistral Large 4 is listed at a sale price (list is $1.36 / $4.18) with no published end date. The -latest aliases are priced the same as the dated id they point to.

Model Input Cached input Cache write Output Batch Notes
Mistral Large 4mistral-large-2610 $0.68 $0.07 — $2.09 50% off Sale price; list is $1.36 / $4.18.
Mistral Large 4mistral-large-latest $0.68 $0.07 — $2.09 50% off Sale price; list is $1.36 / $4.18.
Mistral Large 3legacymistral-large-2512 $0.5 $0.05 — $1.5 50% off
Mistral Medium 3.5mistral-medium-2604 $1.5 $0.15 — $7.5 50% off
Mistral Medium 3.5mistral-medium-latest $1.5 $0.15 — $7.5 50% off
Mistral Small 4mistral-small-2603 $0.15 $0.015 — $0.6 50% off
Mistral Small 4mistral-small-latest $0.15 $0.015 — $0.6 50% off
Ministral 3 14Bministral-14b-2512 $0.2 $0.02 — $0.2 50% off
Ministral 3 14Bministral-14b-latest $0.2 $0.02 — $0.2 50% off
Ministral 3 8Bministral-8b-2512 $0.15 $0.015 — $0.15 50% off
Ministral 3 8Bministral-8b-latest $0.15 $0.015 — $0.15 50% off
Ministral 3 3Bministral-3b-2512 $0.1 $0.01 — $0.1 50% off
Ministral 3 3Bministral-3b-latest $0.1 $0.01 — $0.1 50% off
Codestralcodestral-2508 $0.3 $0.03 — $0.9 50% off
Codestralcodestral-latest $0.3 $0.03 — $0.9 50% off

USD per 1,000,000 tokens, standard tier. A dash under cached input means the provider publishes no cached rate; under cache write, no per-token write charge.

DeepSeek

Source: api-docs.deepseek.com →

Peak rates (01:00 to 04:00 and 06:00 to 10:00 UTC, Monday to Friday). Off-peak is half on every line. Cache hits are billed as cached input. No batch tier.

Model Input Cached input Cache write Output Batch Notes
DeepSeek V4.1 Flashdeepseek-flash $0.3 $0.006 — $1.2 — Off-peak: $0.15 / $0.003 cached / $0.60.
DeepSeek V4 Flash (alias)legacydeepseek-v4-flash $0.3 $0.006 — $1.2 — Retired id, served by V4.1 Flash at Flash prices.
DeepSeek V4 Prodeepseek-v4-pro $1.32 $0.044 — $3.96 — Off-peak: $0.66 / $0.022 cached / $1.98.

USD per 1,000,000 tokens, standard tier. A dash under cached input means the provider publishes no cached rate; under cache write, no per-token write charge.

Cohere

Source: cohere.com →

Cohere publishes no cached-input or batch rate. Command A+ pricing is taken from Cohere's OpenRouter listing because cohere.com does not expose it in HTML; confirm against your invoice. Deprecated models are listed at their existing-customer rates.

Model Input Cached input Cache write Output Batch Notes
Command A+command-a-plus-05-2026 $0.3 — — $1.5 — From Cohere's OpenRouter listing; not shown on cohere.com.
Command Acommand-a-03-2025 $2.5 — — $10 —
Command R+ (08-2024)legacycommand-r-plus-08-2024 $2.5 — — $10 —
Command R (08-2024)legacycommand-r-08-2024 $0.15 — — $0.6 —
Command R7Blegacycommand-r7b-12-2024 $0.0375 — — $0.15 —
Command R+ (04-2024)retiredcommand-r-plus-04-2024 $3 — — $15 — Deprecated September 2025; existing-customer rate.
Command R (03-2024)retiredcommand-r-03-2024 $0.5 — — $1.5 — Deprecated September 2025; existing-customer rate.
Commandretiredcommand $1 — — $2 — Deprecated September 2025; existing-customer rate.
Command Lightretiredcommand-light $0.3 — — $0.6 — Deprecated September 2025; existing-customer rate.

USD per 1,000,000 tokens, standard tier. A dash under cached input means the provider publishes no cached rate; under cache write, no per-token write charge.

Machine-readable: Markdown or JSON, both generated from the same source as this page.

How to read this page

  • Cached input is the price for prompt-prefix tokens served from the provider's cache. It only applies when your prompt repeats; a fresh prompt bills at the input rate.
  • Cache write is Anthropic's one-off premium for storing a prefix (5-minute cache). Google bills cache storage per hour, and OpenAI, xAI, Mistral, DeepSeek and Cohere do not charge for writes.
  • Batch is the discount for asynchronous jobs that can wait up to 24 hours. Agent Ping applies it when a call reports mode: batch.
  • Long prompts cost more on several models (Claude Haiku 5.5 over 100k, Gemini Pro over 200k, every Grok model at 200k, GPT-5.4 and later over 272k). The table lists the base tier; the surcharge is in the notes.
  • Dated ids such as claude-sonnet-4-5-20250929 price the same as their undated row. Agent Ping strips the date when it resolves a rate.
  • Reasoning tokens bill as output on every provider that has them, even though you never see them.

Questions

What does "per million tokens" mean?

Providers bill by token, roughly three quarters of an English word, and quote prices per 1,000,000 of them. Input tokens are what you send (prompt, context, tool results); output tokens are what the model writes back, including hidden reasoning on models that think before answering. Output is usually 4 to 6 times the input price.

What is cached input?

If the start of your prompt is identical to a recent request, most providers serve it from a prompt cache at a steep discount: 0.1x input on most models, 0.05x on Claude Opus 5.5 and Sonnet 5.5, 0.025x on Claude Fable 5.1. Anthropic charges a one-off write premium (1.25x input) to populate the cache; OpenAI does not, and Google bills cache storage by the hour instead.

Which model is cheapest right now?

Of the models on this page, DeepSeek V4.1 Flash off-peak ($0.15 / $0.60), GPT-5 Nano ($0.05 / $0.40), Claude Haiku 5.5 ($0.10 / $0.50 under 100k tokens), GPT-6 Luna ($0.10 / $0.50) and Gemini 2.5 Flash-Lite ($0.10 / $0.40) are the lowest list prices. Cheapest per token is not cheapest per task: a model that needs three retries costs more than one that gets it right.

How often do these prices change?

Every provider on this page changed at least one price in the last six months. We check each provider's own pricing page and update this list and Agent Ping's default rate card together; the verified date at the top is the last check. Scheduled changes we know about (Gemini 4 Argon on 1 December 2026, Gemini 3.6 and 3.8 Flash on 1 January 2027) are in the notes.

How does Agent Ping use this list?

Every LLM call your agents make reports its provider, model and token counts. Agent Ping multiplies them by these rates server-side, per run, so you see cost per agent, per customer and per feature without the SDK ever sending a price. Teams on a negotiated rate override any row in Spend, Rate cards.

Know what each run cost, not just each million.

List prices tell you the rate. Agent Ping applies it to every real run your agents make, so you get cost per agent, per customer and per feature, and a page when one of them spikes.

Try it live in 30 seconds Estimate what your agents cost

Free plan, no card. Prices on this page verified 17 October 2026.