LLM API Pricing Comparison 2026
This page compares text, image, and video generation pricing for every major AI API. We verify prices weekly and flag stale data. Use the built-in calculator to estimate your actual monthly costs based on your input:output ratio and token volume.
Quick Picks
Best Overall
Claude Opus 4.6
Anthropic's most capable model.
$5.00/$25.00 per 1M tokens
Best for Agenting
Claude Opus 4.6
Anthropic's most capable model.
200K context · $25.00/1M out
Best Budget
Kimi K2.5
The dark horse of 2026.
$0.600/$2.50 per 1M tokens
Best ImageGen
FLUX.2 [pro]
Budget-friendly FLUX at just $0.03/MP.
$0.030/image
Compare LLM API Pricing
Showing 35 text models · Last updated 2026-03-08
| Provider ↕ | Model ↕ | Est. Monthly$/mo ↕ | Input $/1M ↕ | Output $/1M ↕ | Cached $/1M ↕ | Context ↕ | ★ ↑ |
|---|---|---|---|---|---|---|---|
| OpenAI | GPT-5.4NEW | $420.00 | $2.50 | $15.00 | $0.250 | 1.1M | 1 |
| OpenAI | GPT-5.2Best Overall | $357.00 | $1.75 | $14.00 | $0.175 | 400K | 1 |
| Anthropic | Claude Opus 4.6Best OverallBest Agenting | $750.00 | $5.00 | $25.00 | $0.500 | 200K | 2 |
| Anthropic | Claude Sonnet 4.6Best OverallBest Agenting | $450.00 | $3.00 | $15.00 | $0.300 | 200K | 3 |
| OpenAI | GPT-5 | $255.00 | $1.25 | $10.00 | $0.125 | 400K | 4 |
Gemini 3.1 ProBest OverallBest Agenting | $336.00 | $2.00 | $12.00 | $0.200 | 1.0M | 6 | |
| OpenAI | GPT-5.3 CodexBest OverallBest Agenting | $357.00 | $1.75 | $14.00 | $0.175 | 400K | 7 |
| Anthropic | Claude Sonnet 4.5 | $450.00 | $3.00 | $15.00 | $0.300 | 200K | 8 |
Gemini 3.1 Flash-LiteNEW | $42.00 | $0.250 | $1.50 | $0.025 | 1.0M | 9 | |
Gemini 2.5 Flash | $63.00 | $0.300 | $2.50 | $0.030 | 1.0M | 10 | |
| DeepSeek | DeepSeek V3.2 ReasonerOSS | $24.36 | $0.280 | $0.420 | $0.028 | 128K | 11 |
| Anthropic | Claude Haiku 4.5 | $150.00 | $1.00 | $5.00 | $0.100 | 200K | 12 |
| OpenAI | GPT-5 mini | $51.00 | $0.250 | $2.00 | $0.025 | 400K | 13 |
Gemini 3 FlashBest Budget | $84.00 | $0.500 | $3.00 | $0.050 | 1.0M | 14 | |
| OpenAI | GPT-5.3 InstantNEW | $357.00 | $1.75 | $14.00 | $0.175 | 400K | 15 |
| DeepSeek | DeepSeek V3.2OSS | $24.36 | $0.280 | $0.420 | $0.028 | 128K | 16 |
| xAI | Grok 4 | $450.00 | $3.00 | $15.00 | $0.750 | 256K | 17 |
| xAI | Grok 4.1 Fast | $21.00 | $0.200 | $0.500 | $0.050 | 2.0M | 18 |
| OpenAI | o4-mini | $145.20 | $1.10 | $4.40 | $0.275 | 200K | 19 |
| Moonshot | Kimi K2.5OSSBest OverallBest BudgetBest Agenting | $81.00 | $0.600 | $2.50 | $0.100 | 262K | 20 |
| Alibaba | Qwen3.5OSSBest Budget | $81.60 | $0.400 | $3.20 | — | 1.0M | 21 |
| Zhipu AI | GLM-5OSSBest Budget | $117.60 | $1.00 | $3.20 | — | 200K | 22 |
| Mistral | Mistral LargeOSS | $57.00 | $0.500 | $1.50 | — | 256K | 23 |
| Mistral | CodestralOSS | $34.20 | $0.300 | $0.900 | — | 256K | 25 |
| MiniMax | Minimax M2.5OSSBest Budget | $39.60 | $0.300 | $1.20 | — | 205K | 26 |
| OpenAI | GPT-5 nano | $10.20 | $0.050 | $0.400 | $0.0050 | 400K | 28 |
| Mistral | Mistral Small 3.2OSS | $11.40 | $0.100 | $0.300 | — | 128K | 31 |
| Anthropic | Claude Opus 4.5 | $750.00 | $5.00 | $25.00 | $0.500 | 200K | 33 |
| Mistral | Mistral Medium 3.1OSS | $60.00 | $0.400 | $2.00 | — | 128K | 36 |
| OpenAI | GPT-5.1 | $255.00 | $1.25 | $10.00 | $0.125 | 400K | 37 |
| OpenAI | GPT-oss-120bOSS | $5.76 | $0.039 | $0.190 | — | 131K | 38 |
| xAI | Grok Code Fast | $39.00 | $0.200 | $1.50 | $0.020 | 256K | 39 |
Gemini 2.5 Pro | $255.00 | $1.25 | $10.00 | $0.125 | 1.0M | 40 | |
| Mistral | Magistral Medium 1.2 | $210.00 | $2.00 | $5.00 | — | 128K | 41 |
| Mistral | Devstral 2OSS | $60.00 | $0.400 | $2.00 | — | 256K | 42 |
Keeping track of LLM API costs in 2026 is a full-time job. OpenAI, Anthropic, Google, DeepSeek, Meta, Mistral, xAI, Moonshot, and more all offer models at wildly different price points — and the pricing pages update faster than most developers can keep up.
The short version: for best-in-class intelligence, Claude Sonnet 4.6 ($3/$15 per 1M) and Gemini 3.1 Pro ($2/$12 per 1M with 1M context) lead the pack. For agentic coding, GPT-5.3 Codex and Kimi K2.5 ($0.60/$2.50, open-weight) are the top picks. On the budget end, Gemini 3 Flash ($0.50/$3.00) and DeepSeek V3 ($0.14/$0.28) deliver incredible value. For image gen, FLUX.2 [pro] at $0.03/img can't be beat. For video, Grok Imagine ($0.05/sec) is the cheapest and Google VEO 3.1 has the best quality. For deeper analysis, see our budget coding LLMs guide and how to pick the right model for your work.
