LIA LitePicks a model for you among the cheapest onesLIAPicks the model for you, balancing quality and costLIA ProPicks a model for you among the most powerful ones
Models
Every model. One subscription.
Text, image, video, music and voice: the best models of each kind in one place, and LIA sends each request to the one that handles it best. We add new ones as soon as they come out.
Not sure which to pick? LIA decides.
Choose LIA and each request goes to the model that handles it best, balancing quality and cost. LIA Lite goes for the models that use the fewest credits, and LIA Pro for the most powerful ones. Or pick any model yourself.
Low, medium or high: what a typical use of each model costs compared with the others of its kind, based on its rate. Each kind is compared on its own: a “low” video uses more credits than a “high” text. Models without a label have no rate to compare.
Text
Chat, reasoning, writing and analysis.
GPT-6.1 Terra, cost mediumClear writing and understanding imagesGPT-6.1 Sol, cost mediumComplex code, analysis and professional workGPT-6 Astra, cost highThe hardest professional and coding workGPT-6 Sol, cost mediumComplex code and multi-step tasksGPT-6 Luna, cost lowFocused tasks in large numbersGPT-5.6 Sol, cost highComplex professional workGPT-5.6 Terra, cost mediumWork that needs a balance of quality and costGPT-5.6 Luna, cost lowTasks in large numbers, spending littleGPT-5.5, cost highComplex professional work and codeGPT-5.5 Pro, cost highHard problems: it thinks longer, sometimes minutesGPT-5.4, cost mediumCode and professional workGPT-5.4 Pro, cost highHard problems: it thinks longer, sometimes minutesGPT-5.4 mini, cost mediumCode and tasks in large numbers, fastGPT-5.4 nano, cost lowSimple tasks in large numbersGPT-5.2, cost mediumComplex professional workGPT-5.2 Pro, cost highHard problems: it thinks longer, sometimes minutesGPT-5.1, cost mediumCode and multi-step tasks
Gemini 4 Argon, cost highComplex analysis and code, with very long documentsGemini 3.8 Flash, cost mediumLong coding tasks, agents and complex workflowsGemini 3.7 Flash, cost mediumEveryday work and code: follows instructions, uses toolsGemini 3.5 Flash, cost mediumCode and multi-step tasks, fastGemini 3.5 Flash-Lite, cost mediumFast tasks in large numbersGemini 3.1 Pro, cost mediumSolving complex problems by reasoningGemini 3.1 Flash-Lite, cost mediumQuick replies in large numbers, such as chatbotsGemini 3 Flash, cost mediumReasoning and multi-step tasks, fastGemini Robotics ER 2, cost mediumRobotics: understanding space and video, planning tasks
Llama 4 Maverick, cost lowChat, creative writing and understanding imagesLlama 4 Scout, cost lowSummarising many documents and large codebases
Qwen 3.8 27B, cost mediumCode, professional work and research; open modelGPT-OSS 120B, cost lowReasoning and tool use; open model
Claude Haiku 5.5, cost lowQuick tasks: sorting, extracting data, customer supportClaude Sonnet 5.5, cost mediumEveryday tasks and bug fixing, fastClaude Fable 5.1, cost highThe most demanding reasoning and codeClaude Opus 5.5, cost highLong coding and analysis work, step by stepClaude Sonnet 5, cost mediumPlans and carries out tasks on its own, with toolsClaude Fable 5, cost highThe hardest reasoning and long tasksClaude Opus 4.8, cost highComplex reasoning and multi-step workClaude Opus 4.7, cost highAdvanced code, long tasks, slides and documentsClaude Opus 4.6, cost highCode, research, documents and spreadsheetsClaude Opus 4.5, cost highCode, multi-step tasks, research and spreadsheetsClaude Sonnet 4.5, cost mediumCode, multi-step tasks and reasoningClaude Haiku 4.5, cost mediumFast replies: chat, customer service, coding help
Mistral Medium 3, cost mediumCode, science and maths, and understanding images
DeepSeek V4.1 Flash, cost lowCode and multi-step tasks
Grok 4.7, cost mediumCode, multi-step tasks and knowledge workGrok 4.6, cost mediumCode, multi-step tasks and knowledge workGrok 4.3, cost mediumFast, following instructions and using toolsGrok 4.20 (Reasoning), cost mediumFast, with reasoning and toolsGrok 4.20 (Non-Reasoning), cost mediumFast answers without reasoning, with tools
Image
Generate and edit images from text or images.
FLUX 1.1 Pro Ultra, cost highHigh-resolution images, up to 4 MPFLUX.1 [schnell], cost lowImages from text, fast; open model
Nano Banana, cost mediumCreates and edits images, fastNano Banana Pro, cost highComplex images and multi-step edits, up to 4KNano Banana 2.1, cost mediumCreates and edits images, with readable text, up to 4KGemini 3.1 Flash Image, cost highCreates and edits images from text or other images, up to 4KGemini 3.1 Flash-Lite Image, cost mediumQuick images and edits at 1K
GPT Image 2.5 Sunburst, cost mediumCreates images and edits them preciselyGPT Image 2.5 Flare, cost mediumEveryday images, fast; it also editsGPT Image 2, cost mediumCreates and edits images, fast
Video
Generate video from text, images or clips.
Luma Ray 2, cost high5-second videos from text or start and end images
Hailuo 02, cost mediumVideos that follow complex instructions, up to 1080pMiniMax Video-01, cost medium6-second 720p videos from text or an image
Kling 2.1 Master, cost high5- or 10-second 1080p videos from text or an imageKling 2.1, cost low5- or 10-second videos from an image, at 720p or 1080pKling 2.0, cost high5- or 10-second 720p videos from text
Veo 3.1 Fast, cost mediumVideos with sound from text or images, up to 1080pVeo 3.1, cost highVideos with sound from text or images, up to 4K
Wan 2.2 Animate, cost lowReplaces a character in a video with one from an image
Seedance 2.0, cost highVideos with sound, up to 15 seconds, from text or imagesSeedance 1.0 Pro, cost mediumMulti-shot videos from text or an image, up to 1080pSeedance 1.0 Lite, cost mediumShort videos from text or an image
Mochi 1, cost lowPhotorealistic videos from text, at 480p; open model
Music & audio
Generate music and sound effects.
Lyria 2, cost mediumInstrumental music from a text description
AudioGen, cost lowSound effects from a text descriptionMusicGen Large, cost mediumMusic from a text description; can also continue a clipMusicGen Melody Large, cost mediumMusic from a text description and a melody you provideMusicGen Stereo Large, cost mediumStereo music from a text descriptionMusicGen Stereo Melody Large, cost mediumStereo music from a text description and a melody
Voice
Natural text-to-speech and voice cloning.
GPT Realtime 2.1 Mini, cost mediumReal-time voice: talks with you and reads text aloudTTS-1, cost lowReads text aloud, fastTTS-1 HD, cost mediumReads text aloud, at higher quality
Speech-to-text
Accurate transcription, in real time.
GPT Transcribe, cost lowAccurate transcription of audio and voice messagesWhisper, cost lowGeneral-purpose speech-to-text
Gemini 3.5 Transcribe, cost lowFast transcription that tells speakers apart
Grok Voice Transcribe 2.0, cost lowTranscription in 38+ languages, even when they switch
Vision
Understand images, screenshots and documents.
GPT-6.1 Terra, cost mediumClear writing and understanding imagesGPT-6.1 Sol, cost mediumComplex code, analysis and professional workGPT-6 Luna, cost lowFocused tasks in large numbersGPT-5.6 Sol, cost highComplex professional workGPT-5.1, cost mediumCode and multi-step tasks
Gemini 4 Argon, cost highComplex analysis and code, with very long documentsGemini 3.1 Pro, cost mediumSolving complex problems by reasoningGemini 3.8 Flash, cost mediumLong coding tasks, agents and complex workflowsGemini 3 Flash, cost mediumReasoning and multi-step tasks, fastGemini Robotics ER 2, cost mediumRobotics: understanding space and video, planning tasks
Claude Haiku 5.5, cost lowQuick tasks: sorting, extracting data, customer supportClaude Sonnet 5.5, cost mediumEveryday tasks and bug fixing, fastClaude Fable 5.1, cost highThe most demanding reasoning and codeClaude Opus 5.5, cost highLong coding and analysis work, step by stepClaude Fable 5, cost highThe hardest reasoning and long tasksClaude Opus 4.8, cost highComplex reasoning and multi-step workClaude Opus 4.7, cost highAdvanced code, long tasks, slides and documentsClaude Opus 4.6, cost highCode, research, documents and spreadsheetsClaude Sonnet 4.5, cost mediumCode, multi-step tasks and reasoning
Grok 4.7, cost mediumCode, multi-step tasks and knowledge workGrok 4.20 (Non-Reasoning), cost mediumFast answers without reasoning, with tools
Code
Write, refactor and explain code.
GPT-6.1 Terra, cost mediumClear writing and understanding imagesGPT-6.1 Sol, cost mediumComplex code, analysis and professional workGPT-6 Astra, cost highThe hardest professional and coding workGPT-6 Sol, cost mediumComplex code and multi-step tasksGPT-6 Luna, cost lowFocused tasks in large numbersGPT-5.6 Sol, cost highComplex professional workGPT-5.6 Terra, cost mediumWork that needs a balance of quality and costGPT-5.6 Luna, cost lowTasks in large numbers, spending littleGPT-5.3 CodexProgramming tasks, step by step
Claude Haiku 5.5, cost lowQuick tasks: sorting, extracting data, customer supportClaude Sonnet 5.5, cost mediumEveryday tasks and bug fixing, fastClaude Fable 5.1, cost highThe most demanding reasoning and codeClaude Opus 5.5, cost highLong coding and analysis work, step by stepClaude Fable 5, cost highThe hardest reasoning and long tasksClaude Opus 4.8, cost highComplex reasoning and multi-step workClaude Opus 4.7, cost highAdvanced code, long tasks, slides and documentsClaude Opus 4.6, cost highCode, research, documents and spreadsheetsClaude Sonnet 4.5, cost mediumCode, multi-step tasks and reasoningClaude Haiku 4.5, cost mediumFast replies: chat, customer service, coding help
Gemini 4 Argon, cost highComplex analysis and code, with very long documentsGemini 3.1 Pro, cost mediumSolving complex problems by reasoningGemini 3.8 Flash, cost mediumLong coding tasks, agents and complex workflowsGemini 3.1 Flash-Lite, cost mediumQuick replies in large numbers, such as chatbotsGemini 3 Flash, cost mediumReasoning and multi-step tasks, fast
DeepSeek V4.1 Flash, cost lowCode and multi-step tasks
Model and provider names are trademarks of their respective owners.
Put your AI to work today
Every model and every tool in one place. Start free, no card needed.