Skip to main content
BenchLM

LLM API Pricing Trends

As of September 30, 2026, the frontier token price index is 16, 84% below its March 2023 base. The median flagship model costs $6.00 per million blended tokens across 35 active constituents, and 40 entries, exits, or price changes landed in the latest snapshot.

Pricing verified

Track how LLM API costs have evolved since GPT-4's launch in March 2023. The update ledger records changes between pricing snapshots; the charts show the longer trend.

Largest recorded cut
o3
$40 → $8 output per 1M · −80% on Jul '26
Cheapest frontier output
Grok 4.3
$2.5/1M output · xAI
Frontier price index
100 → 16
84% below March 2023 · 0% month over month

Latest LLM pricing updates

These rows changed between the previous generated price-index snapshot and the September 30, 2026 snapshot. “Entered” and “exited” describe index membership. Price changes use the index's blended rate: three parts input to one part output.

ModelTierSnapshot update
GPT-5.6 SolfrontierCut from $11.25 to $8.00/1M blended (-28.9%)
GPT-5.6 TerramidCut from $5.63 to $4.50/1M blended (-20%)
GPT-5.6 LunabudgetCut from $2.25 to $0.450/1M blended (-80%)
Gemini 3.6 FlashmidCut from $3.00 to $1.50/1M blended (-50%)
Grok 4.1frontierExited the active index
Claude Opus 5.5frontierEntered at $8.00/1M blended
Claude Fable 5.1frontierEntered at $20.00/1M blended
Grok 4.7frontierEntered at $3.00/1M blended
DeepSeek V4.1 FlashmidEntered at $0.525/1M blended
Gemini 3.8 FlashmidEntered at $1.50/1M blended
Gemini 3.7 FlashmidEntered at $1.50/1M blended
Gemini 3.5 Flash-LitebudgetEntered at $0.850/1M blended
Grok 4.5frontierEntered at $3.00/1M blended
Gemini 3.5 FlashmidEntered at $3.38/1M blended
Grok 4.3frontierEntered at $1.56/1M blended
Claude Opus 4.7frontierEntered at $10.00/1M blended
Claude Opus 4.5frontierEntered at $10.00/1M blended
Gemini 2.5 Flash-LitebudgetEntered at $0.175/1M blended
Gemini 3.1 Flash-LitebudgetEntered at $0.563/1M blended
GPT-5 nanobudgetEntered at $0.138/1M blended
GPT-5 (high)frontierEntered at $3.44/1M blended
GPT-5.2 ProfrontierEntered at $57.75/1M blended
GPT-5.3 InstantfrontierEntered at $4.81/1M blended
GPT-6 AstrafrontierEntered at $20.00/1M blended
GPT-6 SolfrontierEntered at $4.00/1M blended
GPT-6.1 SolfrontierEntered at $4.00/1M blended
GPT-6 LunabudgetEntered at $0.200/1M blended
Kimi 2.6frontierEntered at $1.71/1M blended
Ministral 3 14BbudgetEntered at $0.200/1M blended
Ministral 3 3BbudgetEntered at $0.100/1M blended
Ministral 3 8BbudgetEntered at $0.150/1M blended
Mistral Medium 3midEntered at $0.800/1M blended
Mistral Medium 3.5 128BmidEntered at $3.00/1M blended
o1-profrontierEntered at $262.50/1M blended
o4-minibudgetEntered at $1.93/1M blended
Mistral Small 4budgetEntered at $0.262/1M blended
GPT-5 minibudgetEntered at $0.688/1M blended
o3-profrontierEntered at $35.00/1M blended
DeepSeek V4 Pro 0813frontierEntered at $1.98/1M blended
Claude Sonnet 5.5midEntered at $4.00/1M blended

Snapshot changes can include a corrected registry value or cohort change. Check the live LLM API pricing comparison before changing a production budget.

Frontier token price trend (March 2023 = 100)

Median blended token price across active frontier models, normalized to GPT-4 launch pricing. Lower means cheaper.

Full index, sub-indexes and constituents →

Output price history by model

Output price per million tokens over time. Toggle models to compare.

OpenAIAnthropicGoogleDeepSeekMistralxAI

Radar markers identify recorded price changes in this dated history; they do not represent measured alert lead time or provider announcement timestamps. Radar events attach the published provider source when the change is confirmed.

Current LLM API prices

Per million tokens, highest output price first

Model / providerInput / output
o1-proOpenAI · since Sep '26
$150 input$600 output
GPT-5.4 ProOpenAI · since Apr '26
$30 input$180 output
GPT-5.5 ProOpenAI · since Apr '26
$30 input$180 output
GPT-5.2 ProOpenAI · since Sep '26
$21 input$168 output
o3-proOpenAI · since Sep '26
$20 input$80 output
Claude 3 OpusAnthropic · since Mar '24
$15 input$75 output
GPT-4OpenAI · since Mar '23
$30 input$60 output
o1OpenAI · since Sep '24
$15 input$60 output
Claude Fable 5Anthropic · since Jun '26
$10 input$50 output
Claude Fable 5.1Anthropic · since Sep '26
$10 input$50 output
GPT-6 AstraOpenAI · since Sep '26
$10 input$50 output
GPT-4 TurboOpenAI · since Nov '23
$10 input$30 output
GPT-5.5OpenAI · since Apr '26
$5 input$30 output
Claude Opus 4.6Anthropic · since Feb '26
$5 input$25 output
Claude Opus 4.8Anthropic · since May '26
$5 input$25 output
Claude Opus 5Anthropic · since Jul '26
$5 input$25 output
Claude Opus 4.7Anthropic · since Sep '26
$5 input$25 output
Claude Opus 4.5Anthropic · since Sep '26
$5 input$25 output
Claude Opus 5.5Anthropic · since Sep '26
$4 input$20 output
GPT-5.6 SolOpenAI · since Jul '26
$4 input$20 output
GPT-5.4OpenAI · since Apr '26
$2.5 input$15 output
Claude 3 SonnetAnthropic · since Mar '24
$3 input$15 output
Claude 3.5 SonnetAnthropic · since Jun '24
$3 input$15 output
Claude Sonnet 4.5Anthropic · since Sep '25
$3 input$15 output
Claude Sonnet 4.6Anthropic · since Feb '26
$3 input$15 output
Grok 4.1xAI · since Feb '26
$3 input$15 output
Kimi K3Moonshot AI · since Jul '26
$3 input$15 output
GPT-5.2OpenAI · since Feb '26
$1.75 input$14 output
GPT-5.3 InstantOpenAI · since Sep '26
$1.75 input$14 output
Gemini 3.1 ProGoogle · since Apr '26
$2 input$12 output
GPT-5.6 TerraOpenAI · since Jul '26
$2 input$12 output
GPT-4oOpenAI · since May '24
$2.5 input$10 output
GPT-5.1OpenAI · since Sep '25
$1.25 input$10 output
Gemini 2.5 ProGoogle · since Jun '25
$1.25 input$10 output
Claude Sonnet 5Anthropic · since Jun '26
$2 input$10 output
GPT-5 (high)OpenAI · since Sep '26
$1.25 input$10 output
GPT-6 SolOpenAI · since Sep '26
$2 input$10 output
GPT-6.1 SolOpenAI · since Sep '26
$2 input$10 output
Claude Sonnet 5.5Anthropic · since Sep '26
$2 input$10 output
Gemini 3.5 FlashGoogle · since Sep '26
$1.5 input$9 output
o3OpenAI · since Apr '25
$2 input$8 output
GPT-4.1OpenAI · since Apr '25
$2 input$8 output
Mistral Medium 3.5 128BMistral · since Sep '26
$1.5 input$7.5 output
Grok 4.6xAI · since Aug '26
$2 input$6 output
Grok 4.7xAI · since Sep '26
$2 input$6 output
Grok 4.5xAI · since Sep '26
$2 input$6 output
Claude Haiku 4.5Anthropic · since Sep '25
$1 input$5 output
Gemini 1.5 ProGoogle · since May '24
$1.25 input$5 output
GPT-5.4 miniOpenAI · since Apr '26
$0.75 input$4.5 output
o3-miniOpenAI · since Apr '25
$1.1 input$4.4 output
o4-miniOpenAI · since Sep '26
$1.1 input$4.4 output
Kimi 2.6Moonshot AI · since Sep '26
$0.95 input$4 output
DeepSeek V4 Pro 0813DeepSeek · since Sep '26
$1.32 input$3.96 output
Gemini 3.6 FlashGoogle · since Jul '26
$0.75 input$3.75 output
Gemini 3.8 FlashGoogle · since Sep '26
$0.75 input$3.75 output
Gemini 3.7 FlashGoogle · since Sep '26
$0.75 input$3.75 output
Gemini 3 FlashGoogle · since Dec '25
$0.5 input$3 output
Gemini 2.5 FlashGoogle · since Dec '24
$0.3 input$2.5 output
Gemini 3.5 Flash-LiteGoogle · since Sep '26
$0.3 input$2.5 output
Grok 4.3xAI · since Sep '26
$1.25 input$2.5 output
DeepSeek R1DeepSeek · since Jan '25
$0.55 input$2.19 output
Mistral Medium 3Mistral · since Sep '26
$0.4 input$2 output
GPT-5 miniOpenAI · since Sep '26
$0.25 input$2 output
GPT-4.1 miniOpenAI · since Apr '25
$0.4 input$1.6 output
Mistral Large 3Mistral · since Feb '25
$0.5 input$1.5 output
Gemini 3.1 Flash-LiteGoogle · since Sep '26
$0.25 input$1.5 output
Claude 3 HaikuAnthropic · since Mar '24
$0.25 input$1.25 output
GPT-5.4 nanoOpenAI · since Mar '26
$0.2 input$1.25 output
GPT-5.6 LunaOpenAI · since Jul '26
$0.2 input$1.2 output
DeepSeek V4.1 FlashDeepSeek · since Sep '26
$0.3 input$1.2 output
DeepSeek V3DeepSeek · since Jan '25
$0.27 input$1.1 output
GPT-4o miniOpenAI · since Jul '24
$0.15 input$0.6 output
Mistral Small 4Mistral · since Sep '26
$0.15 input$0.6 output
Grok 3 MinixAI · since Apr '25
$0.3 input$0.5 output
GPT-6 LunaOpenAI · since Sep '26
$0.1 input$0.5 output
GPT-4.1 nanoOpenAI · since Apr '25
$0.1 input$0.4 output
Gemini 2.5 Flash-LiteGoogle · since Sep '26
$0.1 input$0.4 output
GPT-5 nanoOpenAI · since Sep '26
$0.05 input$0.4 output
Gemini 1.5 FlashGoogle · since May '24
$0.075 input$0.3 output
Ministral 3 14BMistral · since Sep '26
$0.2 input$0.2 output
Ministral 3 8BMistral · since Sep '26
$0.15 input$0.15 output
Ministral 3 3BMistral · since Sep '26
$0.1 input$0.1 output

Prices per million tokens from official API pricing. Does not include batch, cached, or volume discounts.

Questions

How much have LLM API prices dropped?

The frontier LLM token price index is 16 as of September 30, 2026, which is 84% below its March 2023 base of 100. The index uses a 3:1 input-to-output blend and the median price across active frontier models, so it measures the cohort rather than one provider's cheapest SKU.

Which LLM provider is cheapest right now?

Grok 4.3 from xAI has the lowest current output rate in this page's selected frontier history set at $2.5/1M tokens. That is not a universal cheapest-model verdict: the live comparison includes more models, while this history page keeps only rows with usable dated snapshots.

Why do reasoning models cost more?

Reasoning models like o1, o3, and DeepSeek R1 generate internal chain-of-thought tokens before producing the final answer. These thinking tokens consume compute but are billed as output tokens, making the effective cost per useful output token significantly higher.

Will LLM prices keep dropping?

There is no guarantee of a straight decline. Frontier token prices remain 84% below March 2023, but the current index moved down 0% month over month as the active model mix and published rates changed. Use the monthly series instead of assuming every new model will be cheaper.

Get the weekly price-change file

Material API price moves, new model rates, and the pricing changes worth adding to your next review.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.