unitcostai

Snapshot Reference

Pricing Snapshot Reference

Current base token rates used by the calculators.

Snapshot date: 2026-10-159 providers129 models

Current snapshot: 2026-10-15. Included providers: OpenAI, Anthropic, Google, xAI, DeepSeek, Kimi, Qwen (International), Mistral, Cohere. Lead rows include copyable input and output prices.

How To Use This Reference

  • Verify the exact base token rates used by the live calculators.
  • Start with the lead models shown first in each provider card.
  • Expand a provider card when you need the full snapshot catalog.

Standard input and output token rates, including published prompt-size tiers where modeled. Provider prompt-cache reads/writes, batch discounts, and search/tool/session fees are excluded.

Methodology and corrections

Daily updates collect prices from the official sources linked below. Calculators use the published snapshot and the inputs you enter. A snapshot date does not verify your workload, model quality, or negotiated rates.

If a source cannot be fetched or parsed, previous rows may be retained; Snapshot Update Notes identify affected providers. Application response caching is modeled separately from provider prompt-cache billing.

To report a pricing error, contact us with the model, displayed rate, and official source. We check the discrepancy against that source before updating the published snapshot.

Snapshot Model Catalog

This is the full model catalog in the current snapshot. Prompt-size tier badges only appear when the snapshot has explicit request-size thresholds.

OpenAI

  • GPT-6 Astra: $10 to $20 input / 1M tokens, $50 to $75 output / 1M tokens.
  • GPT-6.1 Sol: $2 to $4 input / 1M tokens, $10 to $15 output / 1M tokens.
  • GPT-5.6 Terra: $2 to $4 input / 1M tokens, $12 to $18 output / 1M tokens.
ModelInput / 1MOutput / 1M
GPT-6 Astra
$10 to $20$50 to $75
Pricing details
Prompt-size tiers
  • <=272K input tokens: $10 input / $50 output per 1M tokens.
  • >272K input tokens: $20 input / $75 output per 1M tokens.

Published Standard API rates; access depends on OpenAI rollout and account eligibility. Provider prompt-cache reads/writes and optional processing modes are not modeled.

GPT-6.1 Sol
$2 to $4$10 to $15
Pricing details
Prompt-size tiers
  • <=272K input tokens: $2 input / $10 output per 1M tokens.
  • >272K input tokens: $4 input / $15 output per 1M tokens.
GPT-5.6 Terra
$2 to $4$12 to $18
Pricing details
Prompt-size tiers
  • <=272K input tokens: $2 input / $12 output per 1M tokens.
  • >272K input tokens: $4 input / $18 output per 1M tokens.
GPT-6 Luna
$0.1 to $0.2$0.5 to $0.75
Pricing details
Prompt-size tiers
  • <=272K input tokens: $0.1 input / $0.5 output per 1M tokens.
  • >272K input tokens: $0.2 input / $0.75 output per 1M tokens.
GPT-5.5
$5$30
GPT-5.4 Mini
$0.75$4.5

Model

GPT-6 Astra

Input / 1M

$10 to $20

Output / 1M

$50 to $75
Pricing details
Prompt-size tiers
  • <=272K input tokens: $10 input / $50 output per 1M tokens.
  • >272K input tokens: $20 input / $75 output per 1M tokens.

Published Standard API rates; access depends on OpenAI rollout and account eligibility. Provider prompt-cache reads/writes and optional processing modes are not modeled.

Model

GPT-6.1 Sol

Input / 1M

$2 to $4

Output / 1M

$10 to $15
Pricing details
Prompt-size tiers
  • <=272K input tokens: $2 input / $10 output per 1M tokens.
  • >272K input tokens: $4 input / $15 output per 1M tokens.

Model

GPT-5.6 Terra

Input / 1M

$2 to $4

Output / 1M

$12 to $18
Pricing details
Prompt-size tiers
  • <=272K input tokens: $2 input / $12 output per 1M tokens.
  • >272K input tokens: $4 input / $18 output per 1M tokens.

Model

GPT-6 Luna

Input / 1M

$0.1 to $0.2

Output / 1M

$0.5 to $0.75
Pricing details
Prompt-size tiers
  • <=272K input tokens: $0.1 input / $0.5 output per 1M tokens.
  • >272K input tokens: $0.2 input / $0.75 output per 1M tokens.

Model

GPT-5.5

Input / 1M

$5

Output / 1M

$30

Model

GPT-5.4 Mini

Input / 1M

$0.75

Output / 1M

$4.5

Anthropic

  • Claude Fable 5.1: $10 input / 1M tokens, $50 output / 1M tokens.
  • Claude Mythos 5.1: $10 input / 1M tokens, $50 output / 1M tokens.
  • Claude Opus 5.5: $4 input / 1M tokens, $20 output / 1M tokens.
ModelInput / 1MOutput / 1M
Claude Fable 5.1
$10$50
Claude Mythos 5.1
$10$50
Claude Opus 5.5
$4$20
Claude Sonnet 5.5
$2$10
Claude Haiku 5.5
$0.1 to $0.5$0.5 to $2.5
Pricing details
Prompt-size tiers
  • <=100,000 input tokens: $0.1 input / $0.5 output per 1M tokens.
  • >100,000 input tokens: $0.5 input / $2.5 output per 1M tokens.

Model

Claude Fable 5.1

Input / 1M

$10

Output / 1M

$50

Model

Claude Mythos 5.1

Input / 1M

$10

Output / 1M

$50

Model

Claude Opus 5.5

Input / 1M

$4

Output / 1M

$20

Model

Claude Sonnet 5.5

Input / 1M

$2

Output / 1M

$10

Model

Claude Haiku 5.5

Input / 1M

$0.1 to $0.5

Output / 1M

$0.5 to $2.5
Pricing details
Prompt-size tiers
  • <=100,000 input tokens: $0.1 input / $0.5 output per 1M tokens.
  • >100,000 input tokens: $0.5 input / $2.5 output per 1M tokens.

Google

  • Gemini 3.1 Pro Preview: $2 to $4 input / 1M tokens, $12 to $18 output / 1M tokens.
  • Gemini 3.8 Flash: $0.75 input / 1M tokens, $3.75 output / 1M tokens.
  • Gemini 3.8 Flash Lite Tts: $0.5 input / 1M tokens, $6 output / 1M tokens.
ModelInput / 1MOutput / 1M
Gemini 3.1 Pro Preview
$2 to $4$12 to $18
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $2 input / $12 output per 1M tokens.
  • Over 200K prompt tokens / request: $4 input / $18 output per 1M tokens.
Gemini 3.8 Flash
$0.75$3.75
Pricing details
Limited-time rate

Google promotional standard rates apply through December 31, 2026; the source lists $1.50 input and $7.50 output per 1M tokens starting January 1, 2027.

Gemini 3.8 Flash Lite Tts
$0.5$6

Model

Gemini 3.1 Pro Preview

Input / 1M

$2 to $4

Output / 1M

$12 to $18
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $2 input / $12 output per 1M tokens.
  • Over 200K prompt tokens / request: $4 input / $18 output per 1M tokens.

Model

Gemini 3.8 Flash

Input / 1M

$0.75

Output / 1M

$3.75
Pricing details
Limited-time rate

Google promotional standard rates apply through December 31, 2026; the source lists $1.50 input and $7.50 output per 1M tokens starting January 1, 2027.

Model

Gemini 3.8 Flash Lite Tts

Input / 1M

$0.5

Output / 1M

$6

xAI

  • Grok 4.7: $2 to $4 input / 1M tokens, $6 to $12 output / 1M tokens.
  • Grok 4.20 Multi-Agent: $1.25 to $2.5 input / 1M tokens, $2.5 to $5 output / 1M tokens.
  • Grok 4.20 Reasoning: $1.25 to $2.5 input / 1M tokens, $2.5 to $5 output / 1M tokens.
ModelInput / 1MOutput / 1M
Grok 4.7
$2 to $4$6 to $12
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $2 input / $6 output per 1M tokens.
  • Over 200K prompt tokens / request: $4 input / $12 output per 1M tokens.
Grok 4.20 Multi-Agent
$1.25 to $2.5$2.5 to $5
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $1.25 input / $2.5 output per 1M tokens.
  • Over 200K prompt tokens / request: $2.5 input / $5 output per 1M tokens.
Grok 4.20 Reasoning
$1.25 to $2.5$2.5 to $5
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $1.25 input / $2.5 output per 1M tokens.
  • Over 200K prompt tokens / request: $2.5 input / $5 output per 1M tokens.

Model

Grok 4.7

Input / 1M

$2 to $4

Output / 1M

$6 to $12
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $2 input / $6 output per 1M tokens.
  • Over 200K prompt tokens / request: $4 input / $12 output per 1M tokens.

Model

Grok 4.20 Multi-Agent

Input / 1M

$1.25 to $2.5

Output / 1M

$2.5 to $5
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $1.25 input / $2.5 output per 1M tokens.
  • Over 200K prompt tokens / request: $2.5 input / $5 output per 1M tokens.

Model

Grok 4.20 Reasoning

Input / 1M

$1.25 to $2.5

Output / 1M

$2.5 to $5
Pricing details
Prompt-size tiers
  • Up to 200K prompt tokens / request: $1.25 input / $2.5 output per 1M tokens.
  • Over 200K prompt tokens / request: $2.5 input / $5 output per 1M tokens.

DeepSeek

  • DeepSeek V4 Pro: $1.32 input / 1M tokens, $3.96 output / 1M tokens.
  • DeepSeek V4.1 Flash: $0.3 input / 1M tokens, $1.2 output / 1M tokens.
  • DeepSeek Flash (legacy alias): $0.3 input / 1M tokens, $1.2 output / 1M tokens.
ModelInput / 1MOutput / 1M
DeepSeek V4 Pro
$1.32$3.96
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays.

DeepSeek V4.1 Flash
$0.3$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

DeepSeek Flash (legacy alias)
$0.3$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

DeepSeek Chat
$0.3$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

DeepSeek Reasoner
$0.3$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

Model

DeepSeek V4 Pro

Input / 1M

$1.32

Output / 1M

$3.96
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays.

Model

DeepSeek V4.1 Flash

Input / 1M

$0.3

Output / 1M

$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

Model

DeepSeek Flash (legacy alias)

Input / 1M

$0.3

Output / 1M

$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

Model

DeepSeek Chat

Input / 1M

$0.3

Output / 1M

$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

Model

DeepSeek Reasoner

Input / 1M

$0.3

Output / 1M

$1.2
Pricing details
Time-window rates

DeepSeek peak rates are used for deterministic estimates. Off-peak rates are 50% lower. Peak hours are 01:00 - 04:00 and 06:00 - 10:00 UTC, Monday through Friday, excluding Chinese public holidays. DeepSeek V4.1 Flash serves deepseek-flash and the retired deepseek-v4-flash / deepseek-v4-flash-vision-exp aliases.

Kimi

  • Kimi K3: $3 input / 1M tokens, $15 output / 1M tokens.
  • Kimi K2.7 Code: $0.95 input / 1M tokens, $4 output / 1M tokens.
ModelInput / 1MOutput / 1M
Kimi K3
$3$15
Pricing details

Kimi direct API standard USD rates; uncached input and output tokens. Tool charges and cache discounts are excluded.

Kimi K2.7 Code
$0.95$4
Pricing details

Kimi direct API standard USD rates; uncached input and output tokens. Tool charges and cache discounts are excluded.

Model

Kimi K3

Input / 1M

$3

Output / 1M

$15
Pricing details

Kimi direct API standard USD rates; uncached input and output tokens. Tool charges and cache discounts are excluded.

Model

Kimi K2.7 Code

Input / 1M

$0.95

Output / 1M

$4
Pricing details

Kimi direct API standard USD rates; uncached input and output tokens. Tool charges and cache discounts are excluded.

Qwen (International)

  • Qwen 3.8 Max: $2 input / 1M tokens, $6 output / 1M tokens.
  • Qwen 3.8 Flash: $0.15 input / 1M tokens, $0.47 output / 1M tokens.
ModelInput / 1MOutput / 1M
Qwen 3.8 Max
$2$6
Pricing details

Alibaba Cloud Singapore international standard USD rates. Context tiers use decimal K/M tokens. Cache and batch discounts are excluded.

Qwen 3.8 Flash
$0.15$0.47
Pricing details

Alibaba Cloud Singapore international standard USD rates. Context tiers use decimal K/M tokens. Cache and batch discounts are excluded.

Model

Qwen 3.8 Max

Input / 1M

$2

Output / 1M

$6
Pricing details

Alibaba Cloud Singapore international standard USD rates. Context tiers use decimal K/M tokens. Cache and batch discounts are excluded.

Model

Qwen 3.8 Flash

Input / 1M

$0.15

Output / 1M

$0.47
Pricing details

Alibaba Cloud Singapore international standard USD rates. Context tiers use decimal K/M tokens. Cache and batch discounts are excluded.

Mistral

  • Mistral Large 3: $0.5 input / 1M tokens, $1.5 output / 1M tokens.
  • Mistral Medium 3.5: $1.5 input / 1M tokens, $7.5 output / 1M tokens.
  • Mistral Small 26.03: $0.15 input / 1M tokens, $0.6 output / 1M tokens.
ModelInput / 1MOutput / 1M
Mistral Large 3
$0.5$1.5
Mistral Medium 3.5
$1.5$7.5
Mistral Small 26.03
$0.15$0.6
Codestral 25.08
$0.3$0.9

Model

Mistral Large 3

Input / 1M

$0.5

Output / 1M

$1.5

Model

Mistral Medium 3.5

Input / 1M

$1.5

Output / 1M

$7.5

Model

Mistral Small 26.03

Input / 1M

$0.15

Output / 1M

$0.6

Model

Codestral 25.08

Input / 1M

$0.3

Output / 1M

$0.9

Cohere

  • Command A+: $0 input / 1M tokens, $0 output / 1M tokens.
  • Command A: $2.5 input / 1M tokens, $10 output / 1M tokens.
  • Command R+: $2.5 input / 1M tokens, $10 output / 1M tokens.
ModelInput / 1MOutput / 1M
Command A+
$0$0
Pricing details
Usage-limited rate

Cohere lists Command A+ as free for trial and production keys until published rate limits are reached.

Command A
$2.5$10
Command R+
$2.5$10
Command R
$0.15$0.6
Command R7B
$0.0375$0.15

Model

Command A+

Input / 1M

$0

Output / 1M

$0
Pricing details
Usage-limited rate

Cohere lists Command A+ as free for trial and production keys until published rate limits are reached.

Model

Command A

Input / 1M

$2.5

Output / 1M

$10

Model

Command R+

Input / 1M

$2.5

Output / 1M

$10

Model

Command R

Input / 1M

$0.15

Output / 1M

$0.6

Model

Command R7B

Input / 1M

$0.0375

Output / 1M

$0.15

Snapshot Context

Shared Reference Rates

These are the shared defaults carried into the current snapshot for retrieval-related cost components.

  • Embedding Input$0.02 / 1M tokens
  • Rerank Docs$1 / 1K docs

Source Links

Batch updates read these sources, then publish a versioned in-repo snapshot.

  • OpenAI LLM pricing: source
  • Anthropic LLM pricing: source
  • Google LLM pricing: source
  • xAI model pricing: source
  • DeepSeek model pricing: source
  • Mistral model pricing: source
  • Cohere model pricing: source
  • Kimi direct API model pricing: source
  • Qwen international model pricing: source
  • Embedding pricing reference: source
  • Rerank pricing reference: source

Use This Snapshot In The Tools