LLM API pricing comparison
Every chat model OttoPort carries, priced in the same units by the same seller — so GPT, Claude, Gemini, Kimi, GLM, DeepSeek, Grok, and Qwen can actually be compared. Set your token mix below and the table re-ranks.
These are OttoPort's own published rates, not reproductions of each vendor's list price — 54 models from 10 vendors on one rate card, current as of 2026-10-14. Prices are read from the same catalog the gateway meters requests against, so the table cannot drift from what you are billed.
- Models compared
- 54
- Cheapest at 10M / 2M
- DeepSeek V4 Flash
- $1.41 / month
- Dearest at the same mix
- GPT-6 Astra
- $221.58 / month
Cheapest for your workload
There is no single cheapest LLM API, because input and output are billed at different rates. A summarization job and an agent loop rank the same models in different orders — so put your own volumes in.
| Model | ||||
|---|---|---|---|---|
| DeepSeek V4 FlashDeepSeek | $0.101 | $0.202 | 1,050K | $1.41 |
| GPT-6 LunaOpenAI | $0.111 | $0.554 | 1,050K | $2.22 |
| GPT-6 Luna ProOpenAI | $0.111 | $0.554 | 1,050K | $2.22 |
| Claude Haiku 5.5Anthropic | $0.111 | $0.554 | 1,000K | $2.22 |
| DeepSeek V4.1 FlashDeepSeek | $0.155 | $0.498 | 1,050K | $2.55 |
| Qwen3.8 FlashQwen | $0.166 | $0.521 | 1,000K | $2.70 |
| GLM 5.3 FlashZ.ai | $0.166 | $0.554 | 1,310K | $2.77 |
| GPT-5.6 LunaOpenAI | $0.222 | $1.33 | 1,050K | $4.88 |
| GPT-5.4 nanoOpenAI | $0.222 | $1.38 | 400K | $4.98 |
| Gemini 3.1 Flash-LiteGoogle | $0.277 | $1.66 | 1,048.576K | $6.09 |
| Kimi K2.5Moonshot AI | $0.394 | $2.13 | 262K | $8.20 |
| Gemini 3.5 Flash-LiteGoogle | $0.332 | $2.77 | 1,050K | $8.86 |
| Gemini 2.5 FlashGoogle | $0.332 | $2.77 | 1,048.576K | $8.86 |
| Kimi K2 0711Moonshot AI | $0.598 | $2.42 | 131K | $10.82 |
| Kimi K2 ThinkingMoonshot AI | $0.630 | $2.63 | 262K | $11.56 |
| Kimi K2 0905Moonshot AI | $0.630 | $2.63 | 262K | $11.56 |
| GLM 5.2Z.ai | $0.775 | $2.44 | 1,050K | $12.63 |
| Kimi K2.6Moonshot AI | $0.693 | $3.58 | 262K | $14.09 |
| DeepSeek V4 ProDeepSeek | $1.06 | $2.13 | 1,050K | $14.86 |
| Kimi K2.7 CodeMoonshot AI | $0.756 | $3.68 | 262K | $14.92 |
| Gemini 3.8 FlashGoogle | $0.831 | $4.15 | 1,050K | $16.61 |
| Gemini 3.7 FlashGoogle | $0.831 | $4.15 | 1,050K | $16.61 |
| GPT-5.4 miniOpenAI | $0.831 | $4.98 | 400K | $18.27 |
| GLM 5.3Z.ai | $1.22 | $3.82 | 1,310K | $19.84 |
| DeepSeek V4 Pro 0813DeepSeek | $1.25 | $3.73 | 1,000K | $19.96 |
| Claude Haiku 4.5Anthropic | $1.11 | $5.54 | 200K | $22.18 |
| Grok 4.7xAI | $2.22 | $6.65 | 500K | $35.50 |
| Grok 4.6xAI | $2.22 | $6.65 | 500K | $35.50 |
| Grok 4.5xAI | $2.22 | $6.65 | 500K | $35.50 |
| Qwen3.8 MaxQwen | $2.22 | $6.65 | 1,000K | $35.50 |
| Gemini 3.5 FlashGoogle | $1.66 | $9.97 | 1,048.576K | $36.54 |
| GPT-6.1 SolOpenAI | $2.22 | $11.08 | 1,050K | $44.36 |
| GPT-6.1 Sol ProOpenAI | $2.22 | $11.08 | 1,050K | $44.36 |
| GPT-6 SolOpenAI | $2.22 | $11.08 | 1,050K | $44.36 |
| GPT-6 Sol ProOpenAI | $2.22 | $11.08 | 1,050K | $44.36 |
| GPT-5.6 SolOpenAI | $2.22 | $11.08 | 1,050K | $44.36 |
| Claude Sonnet 5.5Anthropic | $2.22 | $11.08 | 1,000K | $44.36 |
| Claude Sonnet 5Anthropic | $2.22 | $11.08 | 1,000K | $44.36 |
| GPT-5.6 TerraOpenAI | $2.22 | $13.29 | 1,050K | $48.78 |
| Gemini 3.1 Pro PreviewGoogle | $2.22 | $13.29 | 1,048.576K | $48.78 |
| Video UnderstandingOttoPort | $2.22 | $13.29 | 1,048.576K | $48.78 |
| GPT-4oOpenAI | $2.77 | $11.08 | 128K | $49.86 |
| GPT-5.2OpenAI | $1.94 | $15.51 | 400K | $50.42 |
| GPT-5.4OpenAI | $2.77 | $16.62 | 1,000K | $60.94 |
| Kimi K3Moonshot AI | $3.32 | $16.62 | 1,050K | $66.44 |
| Claude Opus 5.5Anthropic | $4.43 | $22.15 | 1,000K | $88.60 |
| Claude Opus 5Anthropic | $5.54 | $27.69 | 1,000K | $110.78 |
| Claude Opus 4.8Anthropic | $5.54 | $27.69 | 1,000K | $110.78 |
| Sonar Deep ResearchPerplexity | $2.22 | $46.53 | 128K | $115.26 |
| GPT-5.5OpenAI | $5.54 | $33.23 | 1,000K | $121.86 |
| GPT-6 AstraOpenAI | $11.08 | $55.39 | 1,050K | $221.58 |
| GPT-6 Astra ProOpenAI | $11.08 | $55.39 | 1,050K | $221.58 |
| Claude Fable 5.1Anthropic | $11.08 | $55.39 | 1,000K | $221.58 |
| Claude Fable 5Anthropic | $11.08 | $55.39 | 1,000K | $221.58 |
Image model pricing
Billed per generated image at the model's base resolution. Larger sizes bill at a multiplier on the same rate.
| Model | Vendor | Rate per image |
|---|---|---|
| GPT Image 1 mini | OpenAI | $0.022 |
| GPT Image 2.5 | OpenAI | $0.026 |
| GPT Image 2.5 Flare | OpenAI | $0.026 |
| Qwen Image Layered | Qwen | $0.028 |
| Seedream 4.7 | ByteDance | $0.033 |
| Seedream 4.7 Sequential | ByteDance | $0.033 |
| Qwen Image 3.0 | Qwen | $0.033 |
| FLUX.2 Pro | Black Forest Labs | $0.033 |
| Seedream 5.0 Layers | ByteDance | $0.037 |
| Seedream 5.0 Lite | ByteDance | $0.039 |
| Qwen Image Angles | Qwen | $0.04 |
| GPT Image 1.5 | OpenAI | $0.044 |
| Nano Banana 2 Lite | $0.044 | |
| Qwen Image 3.0 Pro | Qwen | $0.044 |
| Seedream 5.0 | ByteDance | $0.05 |
| Ideogram 4.5 | Ideogram | $0.066 |
| GPT Image 2 | OpenAI | $0.07 |
| Wan 2.7 Pro Image | Alibaba | $0.083 |
| Nano Banana 2.1 | $0.088 | |
| Nano Banana 2 | $0.088 | |
| Midjourney V7 | Midjourney | $0.088 |
| Midjourney V8 | Midjourney | $0.088 |
| Midjourney V8.2 | Midjourney | $0.088 |
| Midjourney V8.1 | Midjourney | $0.095 |
| Nano Banana Pro | $0.165 |
Video model pricing
Billed per second of finished video, so a 5-second clip costs five times the rate below.
Text-to-speech pricing
Billed per character of input text — roughly 150 spoken words per 1,000 characters.
| Model | Vendor | Rate per 1M characters |
|---|---|---|
| Gemini 2.5 Flash TTS | $11.50 | |
| GPT-4o mini TTS | OpenAI | $13.80 |
| ElevenLabs Turbo 2.5 | ElevenLabs | $51.75 |
| MiniMax Speech 2.8 Turbo | MiniMax | $72.79 |
| ElevenLabs v3 | ElevenLabs | $103.50 |
| MiniMax Speech 2.8 HD | MiniMax | $121.32 |
Music model pricing
Billed per second of generated audio.
| Model | Vendor | Rate per second |
|---|---|---|
| Seed-Music | ByteDance | $0.034 |
| Lyria 2 | $0.069 |
LLM API pricing, answered
What the numbers above do and do not mean.
What is the cheapest LLM API?
It depends on your token mix, because input and output are billed at different rates. At 10M input and 2M output tokens a month, the cheapest of the 54 chat models on OttoPort is DeepSeek V4 Flash at $1.41 a month, and the most expensive is GPT-6 Astra at $221.58. Change the mix on this page and the order changes with it. Rates are as of 2026-10-14.
Are these the same prices the model vendors charge?
No, and the distinction matters. These are OttoPort's own published rates — what OttoPort bills you — for 54 chat models from 10 vendors on one rate card. They are not reproductions of each vendor's list price. The value of the table is that every model is priced in the same units, by the same seller, so the comparison is like for like.
Why is output priced higher than input?
Generating tokens costs more compute than reading them, and every major vendor prices the two separately — usually with output several times dearer. That is why a summarization workload (long input, short output) and an agent workload (short input, long output) can rank the same set of models in completely different orders.
How do I compare image, video, and speech model prices?
They are not comparable in tokens, so they are billed in their own units: image models per generated image, video and music models per second of output, and speech models per 1M characters. This page lists each of those in its own table, cheapest first.
Do the prices on this page change?
Yes. Model rates move as vendors reprice, and this page is generated from the same catalog OttoPort meters requests against, so it reflects the current rate card at build time rather than a hand-maintained copy. The rates shown were current on 2026-10-14. The live catalog is at ottoport.ai/models.
Every model in this table, one API
Switching between any two rows is a one-string change — same endpoint, same key, same OpenAI-compatible request. Prepaid credits, no subscription.