AI API price index · prices read 16 Oct 2026
What does an AI API cost?
11 providers, 271 published prices, each read on the provider’s own pricing page on 16 Oct 2026 and linked to it. Ordered by price and nothing else. Current text models run from $0.10 to $50.00 per million output tokens — 500× — and the same open model costs the same at almost every host that sells it.
How to read this index
Different models are different products. A cheap model and an expensive one are not the same thing at two prices, so the long tables below are a map of the market, not a leaderboard. The one like-for-like comparison is the same open-weight model at several hosts, and it has its own table.
List prices, standard tier. Batch, priority, regional and committed prices are on each provider’s page, not ranked here. Where a page shows a promotional and a regular price, the regular one is ranked and the promotion is noted.
The same open model at every host that prices it
6 of 6 open-weight models cost exactly the same per token at every host that lists them on 16 Oct 2026. Where hosts differ, it is the cached-input price, a regional variant, or the discount — not the list price.
| Model | Host | Input | Cached input | Output |
|---|---|---|---|---|
| Whisper Large v31.2× apart | Together AI | $0.09 per audio hour | ||
| Groq | $0.111 per audio hour | |||
| DeepSeek V4.1 Flashsame price everywhere | Fireworks AI | $0.30 | $0.006 | $1.20 |
| Together AI | $0.30 | $0.006 | $1.20 | |
| GLM 5.3same price everywhere | Fireworks AI | $1.40 | $0.26 | $4.40 |
| Mistral AI | $1.40 | $0.14 | $4.40 | |
| Together AI | $1.40 | $0.26 | $4.40 | |
| GLM 5.3 Flashsame price everywhere | Fireworks AI | $0.15 | $0.03 | $0.50 |
| Together AI | $0.15 | $0.03 | $0.50 | |
| gpt-oss-120bsame price everywhere | Fireworks AI | $0.15 | $0.015 | $0.60 |
| Groq | $0.15 | $0.075 | $0.60 | |
| Together AI | $0.15 | — | $0.60 | |
| Kimi K3same price everywhere | Fireworks AI | $3.00 | $0.30 | $15.00 |
| Together AI | $3.00 | $0.30 | $15.00 | |
| MiniMax M3same price everywhere | Fireworks AI | $0.30 | $0.06 | $1.20 |
| Together AI | $0.30 | $0.06 | $1.20 | |
Text prices per 1M tokens. A dash means the host prints no cached-input price for that model.
Every provider at a glance
Each provider's page has its full price list as published, what it gives free, its discounts, and what the page does not say.
| Provider | Prices read | Text output, per 1M tokens |
|---|---|---|
| OpenAImodel maker | 22 | $0.50 – $50.00 |
| Anthropicmodel maker | 13 | $0.50 – $50.00 |
| Google Geminimodel maker | 16 | $0.40 – $12.00 |
| Mistral AImodel maker | 17 | $0.10 – $7.50 |
| Coheremodel maker | 26 | $0.15 – $1.50 |
| Groqopen-model host | 13 | $0.30 – $4.00 |
| Together AIopen-model host | 51 | $0.14 – $15.00 |
| Fireworks AIopen-model host | 29 | $0.20 – $15.00 |
| Replicateopen-model host | 25 | $10.00 |
| Deepgramvoice | 41 | no text model priced |
| ElevenLabsvoice | 18 | no text model priced |
Text models, cheapest output first
85 current text models, standard tier and standard prompt length, per 1M tokens. Long-prompt bands, regional variants, legacy and contact-sales models are on each provider's page.
| Model · provider | Input | Cached input | Output |
|---|---|---|---|
| Ministral 3 3BMistral AI | $0.10 | $0.01 | $0.10 |
| Llama 3 8B Instruct LiteTogether AI | $0.14 | — | $0.14 |
| Command R7BCohere | $0.0375 | — | $0.15 |
| Ministral 3 8BMistral AI | $0.15 | $0.015 | $0.15 |
| Rnj-1 InstructTogether AI | $0.15 | — | $0.15 |
| NVIDIA Nemotron 3.5 Lightning 30B A3BFireworks AI | $0.05 | $0.01 | $0.20 |
| Ministral 3 14BMistral AI | $0.20 | $0.02 | $0.20 |
| Qwen3.5 9BTogether AI | $0.17 | — | $0.25 |
| DeepSeek V4 Flash 0731Together AI | $0.14 | $0.03 | $0.28 |
| GPT OSS 20B (openai/gpt-oss-20b)Groq | $0.075 | $0.037 | $0.30 |
| Qwen2.5 7B Instruct TurboTogether AI | $0.30 | — | $0.30 |
| Gemini 2.5 Flash-LiteGoogle Gemini | $0.10 | $0.01 | $0.40 |
| Qwen3.8 FlashTogether AI | $0.15 | — | $0.47 |
| Claude Haiku 5.5Anthropic | $0.10 | $0.01 | $0.50 |
| gpt-6-lunaOpenAI | $0.10 | $0.01 | $0.50 |
| GLM 5.3 FlashFireworks AI | $0.15 | $0.03 | $0.50 |
| GLM-5.3-FlashTogether AI | $0.15 | $0.03 | $0.50 |
| Command RCohere | $0.15 | — | $0.60 |
| OpenAI GPT OSS 120BFireworks AI | $0.15 | $0.015 | $0.60 |
| GPT OSS 120B (openai/gpt-oss-120b)Groq | $0.15 | $0.075 | $0.60 |
| Mistral Small 4Mistral AI | $0.15 | $0.015 | $0.60 |
| gpt-oss-120BTogether AI | $0.15 | — | $0.60 |
| Qwen3 235B A22B Instruct 2507 FP8 ThroughputTogether AI | $0.20 | — | $0.60 |
| CodestralMistral AI | $0.30 | $0.03 | $0.90 |
| Gemma 4 31BTogether AI | $0.39 | — | $0.97 |
| Llama 3.3 70BTogether AI | $1.04 | — | $1.04 |
| gpt-5.6-lunaOpenAI | $0.20 | $0.02 | $1.20 |
| DeepSeek V4.1 FlashFireworks AI | $0.30 | $0.006 | $1.20 |
| MiniMax M3Fireworks AI | $0.30 | $0.06 | $1.20 |
| DeepSeek V4.1 FlashTogether AI | $0.30 | $0.006 | $1.20 |
| MiniMax M2.7Together AI | $0.30 | $0.06 | $1.20 |
| MiniMax M3Together AI | $0.30 | $0.06 | $1.20 |
| gpt-5.4-nanoOpenAI | $0.20 | $0.02 | $1.25 |
| Cogito v2.1 671BTogether AI | $1.25 | — | $1.25 |
| Qwen3.7-PlusTogether AI | $0.32 | — | $1.28 |
| Gemini 3.1 Flash-LiteGoogle Gemini | $0.25 | $0.025 | $1.50 |
| Muse Glimmer 30BTogether AI | $0.35 | $0.04 | $1.50 |
| Aya Expanse (8B and 32B)Cohere | $0.50 | — | $1.50 |
| Mistral Large 3Mistral AI | $0.50 | $0.05 | $1.50 |
| NVIDIA Nemotron 3 Ultra (Preview)Fireworks AI | $0.60 | $0.12 | $2.40 |
| Gemini 2.5 FlashGoogle Gemini | $0.30 | $0.03 | $2.50 |
| Gemini 3.5 Flash-LiteGoogle Gemini | $0.30 | $0.03 | $2.50 |
| Gemini 3 Flash PreviewGoogle Gemini | $0.50 | $0.05 | $3.00 |
| Qwen3.6-PlusTogether AI | $0.50 | — | $3.00 |
| Qwen3.5-397B-A17BTogether AI | $0.60 | $0.35 | $3.60 |
| Gemini 3.6 FlashGoogle Gemini | $0.75 | $0.075 | $3.75 |
| Gemini 3.8 FlashGoogle Gemini | $0.75 | $0.075 | $3.75 |
| DeepSeek V4 Pro 0813Together AI | $1.32 | $0.13 | $3.96 |
| Qwen/Qwen3.8-27B (qwen/qwen3.8-27b)Groq | $0.80 | — | $4.00 |
| InklingTogether AI | $1.00 | $0.17 | $4.05 |
| Mistral Large 4Mistral AI | $1.36 | $0.14 | $4.18 |
| GLM 5.3Fireworks AI | $1.40 | $0.26 | $4.40 |
| Z.ai GLM 5.3Mistral AI | $1.40 | $0.14 | $4.40 |
| GLM-5.2Together AI | $1.40 | $0.26 | $4.40 |
| GLM-5.3Together AI | $1.40 | $0.26 | $4.40 |
| gpt-5.4-miniOpenAI | $0.75 | $0.075 | $4.50 |
| Qwen3.7-MaxTogether AI | $1.50 | $0.30 | $4.50 |
| Claude Haiku 4.5Anthropic | $1.00 | $0.10 | $5.00 |
| Qwen 3.8 MaxFireworks AI | $2.00 | $0.25 | $6.00 |
| Qwen3.8-2.4T-A95BTogether AI | $2.00 | $0.25 | $6.00 |
| Mistral Medium 3.5Mistral AI | $1.50 | $0.15 | $7.50 |
| Gemini 2.5 ProGoogle Gemini | $1.25 | $0.125 | $10.00 |
| Claude Sonnet 5Anthropic | $2.00 | $0.20 | $10.00 |
| Claude Sonnet 5.5Anthropic | $2.00 | $0.10 | $10.00 |
| Gemini 4 ArgonGoogle Gemini | $2.00 | $0.10 | $10.00 |
| gpt-6-solOpenAI | $2.00 | $0.20 | $10.00 |
| gpt-6.1-solOpenAI | $2.00 | $0.10 | $10.00 |
| deepseek-ai / deepseek-r1Replicate | $3.75 | — | $10.00 |
| Gemini 3.1 Pro PreviewGoogle Gemini | $2.00 | $0.20 | $12.00 |
| gpt-5.6-terraOpenAI | $2.00 | $0.20 | $12.00 |
| gpt-5.4OpenAI | $2.50 | $0.25 | $15.00 |
| Claude Sonnet 4.6Anthropic | $3.00 | $0.30 | $15.00 |
| Ember-1Fireworks AI | $3.00 | $0.30 | $15.00 |
| Kimi K3Fireworks AI | $3.00 | $0.30 | $15.00 |
| Kimi K3Together AI | $3.00 | $0.30 | $15.00 |
| Claude Opus 5.5Anthropic | $4.00 | $0.20 | $20.00 |
| gpt-5.6-solOpenAI | $4.00 | $0.40 | $20.00 |
| Claude Opus 4.8Anthropic | $5.00 | $0.50 | $25.00 |
| Claude Opus 5Anthropic | $5.00 | $0.50 | $25.00 |
| gpt-5.5OpenAI | $5.00 | $0.50 | $30.00 |
| Claude Fable 5Anthropic | $10.00 | $1.00 | $50.00 |
| Claude Fable 5.1Anthropic | $10.00 | $0.25 | $50.00 |
| Claude Mythos 5Anthropic | $10.00 | $1.00 | $50.00 |
| Claude Mythos 5.1Anthropic | $10.00 | $0.25 | $50.00 |
| gpt-6-astraOpenAI | $10.00 | $1.00 | $50.00 |
Speech to text, per audio hour
19 prices from $0.04 to $0.552 an hour of audio — 14×. Per-minute prices are multiplied by 60.
| Model · provider | Per audio hour | As printed |
|---|---|---|
| Whisper Large V3 Turbo (whisper-large-v3-turbo)Groq | $0.04 | $0.04 per hour |
| NVIDIA Nemotron 3 ASR Streaming 0.6BTogether AI | $0.09 | $0.0015 per minute · Per audio minute |
| NVIDIA Parakeet TDT 0.6B v3Together AI | $0.09 | $0.0015 per minute · Per audio minute |
| Whisper Large v3Together AI | $0.09 | $0.0015 per minute · Per audio minute; docs list openai/whisper-large-v3 as 50% off in the Batch API |
| Whisper (whisper-large-v3)Groq | $0.111 | $0.111 per hour |
| Voxtral Mini Transcribe 2Mistral AI | $0.18 | $0.003 per minute · Per minute of audio input; cached $0.0003 /Min |
| Whisper Large v3 (Streaming)Together AI | $0.21 | $0.0035 per minute · Per audio minute |
| Scribe v2ElevenLabs | $0.22 | $0.22 per hour · batch; keyterm prompting +$0.050/hr, transcript editing +$0.070/hr |
| Scribe v2 MedicalElevenLabs | $0.22 | $0.22 per hour · clinical audio; HIPAA-eligible |
| Nova-3 MonolingualDeepgram | $0.258 | $0.0043 per minute · pre-recorded |
| NVIDIA Nemotron 3.5 ASRTogether AI | $0.27 | $0.0045 per minute · Per audio minute (docs list 'NVIDIA Nemotron 3.5 ASR Streaming 0.6B' at $0.0015 — different model string) |
| Whisper LargeDeepgram | $0.288 | $0.0048 per minute · pre-recorded |
| Nova-3 MultilingualDeepgram | $0.312 | $0.0052 per minute · pre-recorded |
| Voxtral Mini Transcribe RealtimeMistral AI | $0.36 | $0.006 per minute · Price from the GA model card (not on the pricing page) |
| Scribe v2 RealtimeElevenLabs | $0.39 | $0.39 per hour · realtime/streaming; keyterm prompting +$0.080/hr, transcript editing +$0.120/hr |
| Flux EnglishDeepgram | $0.462 | $0.0077 per minute · streaming; regular price |
| Nova-3 MonolingualDeepgram | $0.462 | $0.0077 per minute · streaming; regular price |
| Flux MultilingualDeepgram | $0.468 | $0.0078 per minute · streaming |
| Nova-3 MultilingualDeepgram | $0.552 | $0.0092 per minute · streaming; regular price |
Text to speech, per million characters
12 prices from $15.00 to $80.00 per million characters — 5.3×. Per-thousand prices are multiplied by 1,000.
| Model · provider | Per 1M characters | As printed |
|---|---|---|
| Aura-1Deepgram | $15.00 | $0.015 per 1K characters |
| Voxtral TTSMistral AI | $16.00 | $16.00 per 1M characters · Printed as input $0 /M Chars, output $16 /M Chars (= $0.016 per 1K characters) |
| Canopy Labs Orpheus V1 English (canopylabs/orpheus-v1-english)Groq | $22.00 | $22.00 per 1M characters · Preview model; unit is per 1M characters as printed (= $0.022 per 1K characters) |
| Aura-2Deepgram | $30.00 | $0.03 per 1K characters |
| Flash / TurboElevenLabs | $40.00 | $0.04 per 1K characters · same rate printed for every plan; 40,000 character limit |
| v3 ConversationalElevenLabs | $40.00 | $0.04 per 1K characters · same rate printed for every plan |
| v4 TurboElevenLabs | $40.00 | $0.04 per 1K characters · regular price — the only price on the page since the 72%-off promotion ended on 12 Oct (re-read 16 Oct) |
| Canopy Labs Orpheus Arabic Saudi (canopylabs/orpheus-arabic-saudi)Groq | $40.00 | $40.00 per 1M characters · Preview model; unit is per 1M characters as printed (= $0.04 per 1K characters) |
| Flux TTSDeepgram | $45.00 | $0.045 per 1K characters · matching-credit promo up to $500 through Dec 31, 2026 |
| v2 MultilingualElevenLabs | $80.00 | $0.08 per 1K characters · same rate printed for every plan; 10,000 character limit |
| v3ElevenLabs | $80.00 | $0.08 per 1K characters · same rate printed for every plan; 5,000 character limit |
| v4ElevenLabs | $80.00 | $0.08 per 1K characters · regular price — the only price on the page since the 72%-off promotion ended on 12 Oct (re-read 16 Oct) |
Embeddings, per million input tokens
8 named embedding models, $0.02 to $0.20.
| Model · provider | Per 1M tokens | As printed |
|---|---|---|
| text-embedding-3-smallOpenAI | $0.02 | $0.02 per 1M tokens |
| Embed 5 FastCohere | $0.08 | $0.08 per 1M tokens |
| Qwen3 8BFireworks AI | $0.10 | $0.10 per 1M tokens · billed on input tokens only (fireworks.ai/pricing prints '$0.1') |
| Mistral EmbedMistral AI | $0.10 | $0.10 per 1M tokens · Price from the GA model card (not on the pricing page) |
| Embed 5 ProCohere | $0.12 | $0.12 per 1M tokens |
| text-embedding-3-largeOpenAI | $0.13 | $0.13 per 1M tokens |
| Codestral EmbedMistral AI | $0.15 | $0.15 per 1M tokens |
| Gemini Embedding 2Google Gemini | $0.20 | $0.20 per 1M tokens · text input; free of charge on the free tier |
Download the index — free to reuse with credit
All 271 rows — ranked or not, each with the reason — with the source URL and the date read. Published under CC BY 4.0: use it commercially, change it, republish it. The one condition is credit with a link.
Cite as
Macrostack (2026). AI API price index, read 16 Oct 2026. CC BY 4.0. https://www.macrostack.net/compute/ai-api-prices
Or as a link
<a href="https://www.macrostack.net/compute/ai-api-prices">AI API prices from Macrostack</a>, CC BY 4.0
Think a price here is wrong? Tell us with the provider’s page and we correct it the same day.
How the numbers were made
- Read, not measured. Each price was read on the provider’s own pricing page or official pricing documentation on 16 Oct 2026; every provider’s page names the URL. Where a pricing page and the provider’s own docs disagreed, the pricing page is used and the disagreement is written down.
- Units. Published list prices in US dollars: per million tokens for text and embeddings (input, cached input, output), per audio hour for speech-to-text (per-minute prices × 60), per million characters for text-to-speech (per-thousand prices × 1,000). Standard or pay-as-you-go tier, standard prompt-length band. Where a page prints a promotional and a regular price, the regular one is ranked. Batch, priority, regional and committed prices are described, not ranked.
- Ordered by price and nothing else. No provider pays to be here or to move, and there are no affiliate or sponsored links on these pages.
- Nothing is dropped. A price the tables leave out — a long-prompt band, a legacy model, a Growth-plan rate — stays in the download with the reason.
- Kept fresh by the build. The site will not build this index with prices older than 45 days.
Questions
- How much does an AI API cost per million tokens?
- On 16 Oct 2026, the 85 current text models in this index ranged from $0.10 per million output tokens (Ministral 3 3B, Mistral AI) to $50.00 (Claude Fable 5, Anthropic; Claude Fable 5.1, Anthropic; Claude Mythos 5, Anthropic; Claude Mythos 5.1, Anthropic; gpt-6-astra, OpenAI). Those are different models — the range is the spread of the market, not a ranking of quality.
- Is the same open model cheaper at one host than another?
- Rarely, on list price. Of 6 open-weight models priced by two or more hosts on 16 Oct 2026, 6 cost exactly the same per input and output token everywhere. The differences are in cached-input prices, regional variants and batch discounts.
- Does Macrostack measure these prices?
- No. Every number is the provider's own published price, read on its pricing page on the date shown and linked to it. We do not run benchmarks and nobody pays to be on this page or to move on it.
- Can I reuse this data?
- Yes, under CC BY 4.0, commercial use included. The one condition is credit with a link: Macrostack (2026). AI API price index, read 16 Oct 2026. CC BY 4.0. https://www.macrostack.net/compute/ai-api-prices