llmrelay
// models

Every model, half the price.

Same API. Same models. 50% off official list price on every one. One prepaid balance covers everything.

This list moves. New models get added as they ship, older ones stay listed once we stop pushing them, and occasionally an upstream runs short on a specific model. Anything not callable right now is marked unavailable and shows no price, so you never sign up for something we cannot serve today.

Anthropic Claude

Claude Opus 5
−50%
Current flagship Opus. A generation past 4.8, same list price.
in $2.5/M out $12.5/M 1000K ctx
Claude Sonnet 5
−50%
The next-gen Sonnet. Priced to replace GPT for daily work.
in $1/M out $5/M 1000K ctx
Claude Haiku 4.5
−50%
The fastest and cheapest Claude. Perfect for classification and routing.
in $0.5/M out $2.5/M 200K ctx
Claude Fable 5.1
−50%
Anthropic’s current top Claude. Same $10/$50 list, cheaper cache reads.
in $5/M out $25/M 1000K ctx
Claude Opus 4.6
prev gen
The prior Opus point release. Same price tier as 4.7 and 4.8.
in $2.5/M out $12.5/M 1000K ctx
Claude Opus 4.7
prev gen
Anthropic's flagship reasoning model - the coding and agent workhorse.
in $2.5/M out $12.5/M 1000K ctx
Claude Opus 4.8
prev gen
Previous-generation Opus. Superseded by Opus 5, still served.
in $2.5/M out $12.5/M 1000K ctx
Claude Fable 5
prev gen
Previous Fable. Same $10/$50 list, still served. Use 5.1 for new work.
in $5/M out $25/M 1000K ctx
Claude Sonnet 4.6
unavailable
The everyday driver. 60% of Opus quality at 20% of the cost.
in $1.5/M out $7.5/M 1000K ctx

Google Gemini

Gemini 3.5 Flash
−50%
Fast and cheap, with the full 1M context window.
in $0.75/M out $4.5/M 1049K ctx
Gemini 3.1 Pro
unavailable
Google’s flagship. 1M context, native multimodal.
in $1/M out $6/M 1049K ctx

OpenAI GPT

GPT-6 Astra
−50%
OpenAI’s current flagship. API, ChatGPT paid plans, half list here.
in $5/M out $25/M 1050K ctx
GPT-5.6 Sol
−50%
The full-size GPT-5.6. Frontier reasoning, priced like a mid-tier.
in $2.5/M out $15/M 1050K ctx
GPT-5.6 Terra
−50%
The mid-size GPT-5.6. Half the price of Sol, most of the capability.
in $1.25/M out $7.5/M 1050K ctx
GPT-5.6 Luna
−50%
The small GPT-5.6. Cheap enough for high-volume routing and classification.
in $0.5/M out $3/M 1050K ctx
GPT Image 2
−50%
OpenAI’s current image generation model, billed per token.
prompt $2.5/M image $15/M image gen
GPT-5.5
prev gen
The previous OpenAI flagship. Deep reasoning, top-tier price.
in $2.5/M out $15/M 1050K ctx
GPT-5.4
prev gen
The previous GPT flagship. Still strong, now cheap.
in $1.25/M out $7.5/M 1050K ctx
GPT-5.4 Mini
prev gen
The cheapest GPT we sell. Built for volume.
in $0.375/M out $2.25/M 400K ctx
GPT-5
unavailable
OpenAI's flagship. Sharper reasoning, higher price.
in $2.5/M out $10/M 400K ctx

Other

Grok 4.6
−50%
The latest xAI reasoning model. Successor to Grok 4.5.
in $1/M out $3/M 500K ctx
DeepSeek V4.1 Flash
−50%
Current DeepSeek Flash. Official GA 10 September 2026. Flat $0.19/$0.54 here, no peak surcharge.
in $0.19/M out $0.54/M 1000K ctx
Grok 4.5
prev gen
xAI reasoning model with a 500K context window.
in $1/M out $3/M 500K ctx
DeepSeek V4 Flash
prev gen
Previous Flash id. Still callable here. DeepSeek retired it; new work should use V4.1 Flash.
in $0.19/M out $0.54/M 1000K ctx
Kimi K3
unavailable
Moonshot flagship for long-context coding and knowledge work. 1M context, half the published list.
in $1.5/M out $7.5/M 1049K ctx