llmprices

Compare LLM API prices across providers from the command line.

Web: www.llmcost.run

Find the cheapest model for your prompt in seconds

$ llm-cost calc "Build a Python REST API" --output 500

╭── Cost estimate · 7 input + 500 output ──────────────────────────────────────╮
│  #  Provider      Model                Total cost    vs cheapest              │
│  1  Mistral AI    Mistral Small 3.2    $0.000090     cheapest                 │
│  2  DeepSeek      DeepSeek V4 Flash    $0.000141     1.6x                     │
│  3  Google        Gemini 2.5 Flash-L   $0.000200     2.2x                     │
│  4  xAI           Grok 4.1 Fast        $0.000251     2.8x                     │
│  5  OpenAI        GPT-5.4 Nano         $0.000626     7.0x                     │
│  6  Anthropic     Claude Haiku 4.5     $0.002507     27.9x                    │
│  7  Google        Gemini 3.1 Pro       $0.006014     66.8x                    │
│  8  OpenAI        GPT-5.5              $0.015035    167.1x                    │
╰──────────────────────────────────────────────────────────────────────────────╯

  Cheapest: Mistral Small 3.2 (Mistral AI) — $0.000090

---

Screenshots

---

Install

pip install llmprices

For accurate token counting (uses tiktoken):

pip install "llmprices[tiktoken]"

---

Usage

List all models

llm-cost list

Filter by provider:

llm-cost list --provider anthropic
llm-cost list --provider openai

Filter by efficiency tier:

llm-cost list --tier flagship    # Top-tier models (GPT-5.5, Claude Opus 4.7, o3)
llm-cost list --tier advanced    # Advanced models (GPT-5.4, Claude Sonnet, Gemini Pro)
llm-cost list --tier standard    # Standard models (GPT-5, Claude Haiku, Gemini Flash)
llm-cost list --tier budget      # Budget models (Nano, Small, Lite models)

Sort options (input, output, context, name, value):

llm-cost list --sort output
llm-cost list --sort value       # Sort by efficiency/cost ratio

Search by name:

llm-cost list --search gpt-5
llm-cost list --search gemini

---

Calculate cost for a prompt

# Auto-estimate tokens from text
llm-cost calc "Build a Python REST API" --output 800

# Specify tokens directly
llm-cost calc --input 4000 --output 1000

# Top 5 cheapest only
llm-cost calc --input 10000 --output 2000 --top 5

# Filter to one provider
llm-cost calc "Build a Python REST API" --output 500 --provider google

# Filter by efficiency tier
llm-cost calc "Build a Python REST API" --output 1500 --tier advanced

# Sort by value (efficiency/cost ratio) instead of just cost
llm-cost calc "Build a Python REST API" --output 1000 --sort value --top 10

# One specific model
llm-cost calc "Build a Python REST API" --output 500 --model gpt-5-5

Understanding Value Score: The value score represents the efficiency-to-cost ratio. Higher scores mean better value:

• Budget models often have high value scores for simple tasks
• Advanced/Flagship models have lower value scores but better quality
• Use --sort value to find the best balance for your use case

---

Compare specific models

# Latest flagships head-to-head
llm-cost compare gpt-5-5 claude-opus-4-7 gemini-3-1-pro

# Compare different tiers to see value differences
llm-cost compare gpt-5-5 deepseek-r1 mistral-small-3-2 --input 5000 --output 2000

# Mid-tier sweet spot
llm-cost compare gpt-5-4 claude-sonnet-4-6 gemini-3-flash --input 5000 --output 1000

# Budget tier
llm-cost compare gpt-5-4-nano deepseek-v4-flash grok-4-1-fast mistral-small-3-2

# New agentic models
llm-cost compare deepseek-v4-pro glm-5-1 kimi-k2-6 minimax-m2-7 --input 5000 --output 1000

# From a real prompt
llm-cost compare gpt-5-5 claude-opus-4-7 --prompt "Build a Python REST API"

The comparison table shows:

• Tier: Efficiency tier (flagship/advanced/standard/budget)
• Value: Efficiency-to-cost ratio (higher = better value)
• Total Cost: Complete cost for the specified tokens

---

List providers

llm-cost providers

---

Supported models (May 2026)

Prices in USD per 1M tokens.

Efficiency Tiers

Models are categorized by their capabilities and efficiency:

• Flagship: Top-tier models with maximum efficiency for complex tasks (GPT-5.5, Claude Opus 4.7, o3)
• Advanced: Excellent balance of quality and cost (GPT-5.4, Claude Sonnet, DeepSeek R1, Gemini Pro)
• Standard: Solid performance for most tasks (GPT-5, Claude Haiku, Gemini Flash)
• Budget: Cost-effective for simple tasks (Nano, Small, Lite models)

Provider	Model	Tier	Input	Output	Context
OpenAI	GPT-5.5	Flagship	$5.00	$30.00	1M
OpenAI	GPT-5.5 Pro	Flagship	$30.00	$180.00	1M
OpenAI	GPT-5.4 Pro	Flagship	$30.00	$180.00	400K
OpenAI	o3	Flagship	$10.00	$40.00	200K
Anthropic	Claude Opus 4.7	Flagship	$5.00	$25.00	1M
Anthropic	Claude Opus 4.6	Flagship	$5.00	$25.00	1M
OpenAI	GPT-5.4	Advanced	$2.50	$15.00	1.05M
OpenAI	o4 Mini	Advanced	$1.10	$4.40	200K
Anthropic	Claude Sonnet 4.6	Advanced	$3.00	$15.00	1M
Google	Gemini 3.1 Pro	Advanced	$2.00	$12.00	1M
Google	Gemini 2.5 Pro	Advanced	$1.25	$10.00	1M
xAI	Grok 4	Advanced	$3.00	$15.00	2M
DeepSeek	DeepSeek R1	Advanced	$0.55	$2.19	1M
OpenAI	GPT-5	Standard	$1.25	$10.00	400K
OpenAI	GPT-5.4 Mini	Standard	$0.75	$4.50	400K
Anthropic	Claude Haiku 4.5	Standard	$1.00	$5.00	200K
Google	Gemini 3 Flash	Standard	$0.50	$3.00	1M
Google	Gemini 2.5 Flash	Standard	$0.30	$2.50	1M
DeepSeek	DeepSeek V4 Pro	Standard	$1.74	$3.48	1M
Mistral AI	Mistral Large 3	Standard	$0.50	$1.50	256K
Mistral AI	Mistral Medium 3.5	Standard	$1.00	$3.00	256K
Z.AI	GLM-5.1	Standard	$1.40	$4.40	200K
Kimi	Kimi K2.6	Standard	$0.95	$4.00	256K
Cohere	Command R+	Standard	$3.00	$15.00	128K
OpenAI	GPT-5.4 Nano	Budget	$0.20	$1.25	200K
Google	Gemini 2.5 Flash-Lite	Budget	$0.10	$0.40	1M
xAI	Grok 4.1 Fast	Budget	$0.20	$0.50	2M
DeepSeek	DeepSeek V4 Flash	Budget	$0.14	$0.28	1M
MiniMax	MiniMax M2.7	Budget	$0.30	$1.20	197K
Mistral AI	Mistral Small 3.2	Budget	$0.06	$0.18	131K
Meta	Llama 4 Maverick	Budget	$0.27	$0.85	1M
Meta	Llama 3.3 70B	Budget	$0.59	$0.79	128K
Cohere	Command R7B	Budget	$0.04	$0.15	128K

Notes:

• DeepSeek V4 has two API variants: deepseek-v4-flash and deepseek-v4-pro.
• Cached-input, batch, promotional, long-context, and subscription-plan discounts are not included in the main table.

Price sources:

• DeepSeek: Models & Pricing
• Z.AI: Pricing
• Kimi: Kimi K2.6 Pricing
• MiniMax: Pay as You Go

33 models across 11 providers. Prices stored in llm_cost/data/prices.yaml — PRs to update them are always welcome!

PyPI: pypi.org/project/llmprices

---

Token counting

Default: word-based heuristic — zero extra dependencies. For accurate counts:

pip install "llmprices[tiktoken]"

---

Contributing

The easiest contribution is updating llm_cost/data/prices.yaml when a provider changes their pricing. Each entry is just 4 fields:

my-new-model:
  name: My New Model
  input: 1.50      # $ per 1M input tokens
  output: 6.00     # $ per 1M output tokens
  context: 200000  # context window in tokens

git clone https://github.com/madeburo/llmcost
cd llmprices
pip install -e ".[dev]"
pytest

Web: www.llmcost.run MIT