01
OpenAI API Lowest estimate
GPT-4o mini
$3.08/ month equivalent
COMPARE / LLM APIS
LLM APIs, with the math.
Compare token pricing, prompt caching savings, and inference costs across foundation models.
Commercial API inference for generative language and reasoning models. Shared multi-tenant API endpoints, standard rate limits, and pay-as-you-go billing without committed use discounts or provisioned throughput. Fine-tuning, batch API discounts, embeddings, and audio generation are excluded.
YOUR COMPARISON, YOUR CALL
3 / 4 selectedBring your shortlist.
Choose up to four products for a closer look. The full cost ranking stays below.
Showing 4 of 4 supported products. Find products for this comparison
Your workload
Monthly usage. Adjust the assumptions to match your product.
tokens / month
Uncached input prompt tokens processed by the model API.
Maximum 10,000,000,000 tokens / monthtokens / month
Generated completion and reasoning tokens returned by the model API.
Maximum 2,000,000,000 tokens / monthtokens / month
Prompt tokens read from context cache (system prompts, docs).
Maximum 10,000,000,000 tokens / monthEstimates are in USD before taxes. Your usage stays in this URL.
| Decision guideOne shared workload. Different trade-offs. | |||
|---|---|---|---|
| Monthly equivalentUSD · before tax | Not availableDeepSeek-R1 | $3.08Lowest in your shortlistGPT-4o mini | Not availableClaude 3.5 Haiku |
| Payment commitment | Monthly | Monthly | Monthly |
| Fits this workload? | Outside modeled limits | Within modeled limits | Outside modeled limits |
| What to know | Current DeepSeek pricing lists V4 models. This legacy R1 model and its historical token rates are not supported by the current pricing source. 1 more conditions
| Automatic 50% prompt caching discount applies to prompts with 1,024+ identical leading tokens. | Claude 3.5 Haiku retired on February 19, 2026. This model is unavailable on the Claude API. 1 more conditions
|
| Outside the estimate | This pricing model is not supported. | Fine-tuning, Batch API discounts, audio/realtime APIs, and image generation. | This pricing model is not supported. |
| Evidence | Official pricing ↗Checked 2026-10-08 | Official pricing ↗Checked 2026-10-06 | Official pricing ↗Checked 2026-10-08 |
Look ahead, before you commit.
Vary standard input tokens. Keep all other inputs fixed.
OpenAI API
View exact values and assumptions
Each point follows your selected plan. Automatic choices use the lowest eligible supported plan per product. Missing values mean no supported plan fits. This varies one input, not all costs with customer count. Lines connect sampled estimates; prices between those points can change in steps or at plan limits.
| Standard input tokens | OpenAI API |
|---|---|
| 1,000,000 tokens / month | $1.73 |
| 1,000,001 tokens / month | $1.73 |
| 4,500,000 tokens / month | $2.25 |
| 8,000,000 tokens / month | $2.78 |
| 10,000,000 tokens / month | $3.08 |
| 11,500,000 tokens / month | $3.30 |
| 15,000,000 tokens / month | $3.83 |
| 18,500,000 tokens / month | $4.35 |
| 22,000,000 tokens / month | $4.88 |
| 25,500,000 tokens / month | $5.40 |
| 29,000,000 tokens / month | $5.93 |
| 32,500,000 tokens / month | $6.45 |
| 36,000,000 tokens / month | $6.98 |
| 39,500,000 tokens / month | $7.50 |
| 43,000,000 tokens / month | $8.03 |
| 46,500,000 tokens / month | $8.55 |
| 50,000,000 tokens / month | $9.07 |
| 53,500,000 tokens / month | $9.60 |
| 60,500,000 tokens / month | $10.65 |
| 64,000,000 tokens / month | $11.18 |
| 67,500,000 tokens / month | $11.70 |
| 71,000,000 tokens / month | $12.23 |
| 74,500,000 tokens / month | $12.75 |
| 78,000,000 tokens / month | $13.28 |
| 81,500,000 tokens / month | $13.80 |
| 85,000,000 tokens / month | $14.33 |
| 88,500,000 tokens / month | $14.85 |
| 92,000,000 tokens / month | $15.38 |
| 95,500,000 tokens / month | $15.90 |
| 99,000,000 tokens / month | $16.43 |
| 102,500,000 tokens / month | $16.95 |
| 106,000,000 tokens / month | $17.48 |
| 109,500,000 tokens / month | $18.00 |
| 113,000,000 tokens / month | $18.52 |
| 120,000,000 tokens / month | $19.58 |
Your monthly comparison
2 eligible products in this loaded set · Lowest modeled cost first
02
Llama 3.3 70B (Groq)
$10.43/ month equivalent
2 plans outside this workload
Anthropic Claude · Claude 3.5 Haiku
Claude 3.5 Haiku retired on February 19, 2026. This model is unavailable on the Claude API. Claude 3.5 Sonnet retired on October 28, 2025. This model is unavailable on the Claude API.
DeepSeek API · DeepSeek-R1
Current DeepSeek pricing lists V4 models. This legacy R1 model and its historical token rates are not supported by the current pricing source. Current DeepSeek pricing lists V4 models. This legacy V3 model and its historical token rates are not supported by the current pricing source.
Prices reflect dated, supported plans. Matching a category does not mean matching features. How estimates work
BUYER HUBS & RANKINGS
2026 GuidesLLM APIs Decision Guides
🏆 Choosing LLM APIsTop Picks
Compare modeled costs, documented plan features, and limitations before choosing a provider.
Explore best picks ↗💰 Cheapest LLM APIs LeaderboardLowest Cost
Eligible monthly estimates at a stated reference workload, with sourced charges and exclusions you can inspect.
View cost leaderboard ↗HEAD-TO-HEAD SHOWDOWNS
Detailed ShowdownsPopular LLM APIs Comparisons
OpenAI API vs Anthropic ClaudeShowdown
Compare OpenAI API and Anthropic Claude pricing, limits, and capabilities.
Compare head-to-head ↗OpenAI API vs DeepSeek APIShowdown
Compare OpenAI API and DeepSeek API pricing, limits, and capabilities.
Compare head-to-head ↗OpenAI API vs Groq CloudShowdown
Compare OpenAI API and Groq Cloud pricing, limits, and capabilities.
Compare head-to-head ↗Anthropic Claude vs DeepSeek APIShowdown
Compare Anthropic Claude and DeepSeek API pricing, limits, and capabilities.
Compare head-to-head ↗Anthropic Claude vs Groq CloudShowdown
Compare Anthropic Claude and Groq Cloud pricing, limits, and capabilities.
Compare head-to-head ↗DeepSeek API vs Groq CloudShowdown
Compare DeepSeek API and Groq Cloud pricing, limits, and capabilities.
Compare head-to-head ↗