Undominated.ai

The model landscape

AI model families.
Choose the exact model.

A familiar name can hide a different checkpoint, price tier or product. Explore generations and collections with current catalogue facts and source-scoped notes.

52 families & collections · 318 / 358 standard rows grouped · research reviewed

52 of 52 collections

Prices: listed input / 1M tokens. Frontier: global balanced view.

  • Googlegeneration

    Gemini 3 family

    Compare the Gemini text, preview and image rows separately. Flash is a product tier, and an image-generation row does not become a text substitute because its name starts with Gemini.

    Standard rows
    15
    Rated
    7 / 15
    Input $/M
    $0.250–$2.00
    On global frontier
    1
    8 free / batch variantsSource notes available
  • Z.ailineage

    Z.ai GLM family

    GLM spans general, vision and Flash-labelled models. A shared base or family name does not make post-training, reasoning controls or subscription entitlements identical.

    Standard rows
    16
    Rated
    13 / 16
    Input $/M
    $0.060–$2.80
    On global frontier
    1
    2 free / batch variantsSource notes available
  • NVIDIAportfolio

    NVIDIA Nemotron family

    Lightning, Nano, Super, Ultra and content-safety checkpoints do different jobs. The Nemotron-Flash research family is not evidence of a priced Flash endpoint in this catalogue.

    Standard rows
    5
    Rated
    0 / 5
    Input $/M
    $0.050–$0.625
    On global frontier
    0
    5 free / batch variantsSource notes available
  • StepFunlineage

    StepFun Step family

    Step 3.5 Flash and Step 3.7 Flash are distinct checkpoints. Compare observed input modalities, model limits and rating coverage before choosing a replacement.

    Standard rows
    2
    Rated
    1 / 2
    Input $/M
    $0.100–$0.200
    On global frontier
    0
    0 free / batch variantsSource notes available
  • Googlegeneration

    Gemini 2.5 family

    The Gemini 2.5 generation remains separate from Gemini 3. Google limits access to previous active users; that access restriction is not an announced retirement.

    Standard rows
    5
    Rated
    2 / 5
    Input $/M
    $0.100–$1.25
    On global frontier
    0
    3 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-5.1 family

    GPT-5.1 variants retain their own model IDs, prices and rating coverage. API usage is separate from a ChatGPT subscription; Codex and Pro labels need endpoint-specific interpretation.

    Standard rows
    4
    Rated
    1 / 4
    Input $/M
    $0.250–$1.25
    On global frontier
    0
    1 free / batch variantsSource notes available
  • Mistrallineage

    Ministral family

    Ministral generation and revision labels matter. The older Ministral 8B row remains documented as missing; it has not been silently relabelled as Ministral 3.

    Standard rows
    3
    Rated
    0 / 3
    Input $/M
    $0.100–$0.200
    On global frontier
    0
    1 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-5 family

    GPT-5 variants retain their own model IDs, prices and rating coverage. API usage is separate from a ChatGPT subscription; Codex and Pro labels need endpoint-specific interpretation.

    Standard rows
    6
    Rated
    3 / 6
    Input $/M
    $0.050–$15.00
    On global frontier
    0
    4 free / batch variantsSource notes available
  • MiniMaxportfolio

    MiniMax M family

    Compare M3, earlier M-series checkpoints and specialised MiniMax products separately. The current M3 model card publishes architecture details and its own community licence; a family-wide MIT assumption is unsafe.

    Standard rows
    8
    Rated
    5 / 8
    Input $/M
    $0.200–$0.400
    On global frontier
    0
    0 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-6.1 family

    Sol is the GPT-6.1 model OpenAI documents. Its Pro-labelled catalogue row is retained separately and is not asserted to be a direct API model ID: OpenAI documents no gpt-6.1-sol-pro model.

    Standard rows
    2
    Rated
    1 / 2
    Input $/M
    $2.00–$2.00
    On global frontier
    0
    0 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-6 family

    Astra, Sol and Luna retain separate catalogue identities, as do their Pro-labelled rows. OpenAI’s direct API reasoning mode and marketplace model slugs are different integration surfaces.

    Standard rows
    6
    Rated
    3 / 6
    Input $/M
    $0.100–$10.00
    On global frontier
    0
    6 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-5.6 family

    Sol, Terra and Luna are distinct named model rows. Pro reasoning mode and reasoning effort are separate controls; a marketplace suffix is not automatically a direct API model ID.

    Standard rows
    6
    Rated
    3 / 6
    Input $/M
    $0.200–$4.00
    On global frontier
    0
    6 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-5.5 family

    GPT-5.5 variants retain their own model IDs, prices and rating coverage. API usage is separate from a ChatGPT subscription; Codex and Pro labels need endpoint-specific interpretation.

    Standard rows
    2
    Rated
    1 / 2
    Input $/M
    $5.00–$30.00
    On global frontier
    0
    2 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-5.4 family

    GPT-5.4 variants retain their own model IDs, prices and rating coverage. API usage is separate from a ChatGPT subscription; Codex and Pro labels need endpoint-specific interpretation.

    Standard rows
    5
    Rated
    3 / 5
    Input $/M
    $0.200–$30.00
    On global frontier
    0
    4 free / batch variantsSource notes available
  • OpenAIcoding

    GPT-5.3 Codex family

    This coding-labelled generation is documented as its own catalogue collection. Compare the exact Codex row and its endpoint constraints rather than assuming a general chat equivalent.

    Standard rows
    1
    Rated
    0 / 1
    Input $/M
    $1.75–$1.75
    On global frontier
    0
    0 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-5.2 family

    GPT-5.2 variants retain their own model IDs, prices and rating coverage. API usage is separate from a ChatGPT subscription; Codex and Pro labels need endpoint-specific interpretation.

    Standard rows
    4
    Rated
    1 / 4
    Input $/M
    $1.75–$21.00
    On global frontier
    0
    2 free / batch variantsSource notes available
  • OpenAIlineage

    OpenAI o series family

    The o-series contains separate reasoning models and dated revisions. Visible output, billed reasoning and the available endpoint controls must remain distinct.

    Standard rows
    8
    Rated
    5 / 8
    Input $/M
    $1.10–$150
    On global frontier
    0
    3 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-4.1 family

    GPT-4.1 variants retain their own model IDs, prices and rating coverage. API usage is separate from a ChatGPT subscription; Codex and Pro labels need endpoint-specific interpretation.

    Standard rows
    3
    Rated
    3 / 3
    Input $/M
    $0.100–$2.00
    On global frontier
    0
    3 free / batch variantsSource notes available
  • OpenAIgeneration

    GPT-4o family

    GPT-4o variants retain their own model IDs, prices and rating coverage. API usage is separate from a ChatGPT subscription; Codex and Pro labels need endpoint-specific interpretation.

    Standard rows
    6
    Rated
    3 / 6
    Input $/M
    $0.150–$5.00
    On global frontier
    0
    2 free / batch variantsSource notes available
  • OpenAIhistorical collection

    GPT-4 and GPT-3.5 family

    These older GPT catalogue rows are retained for existing integrations. Check the direct platform’s lifecycle notice and the actual reseller endpoint independently.

    Standard rows
    6
    Rated
    1 / 6
    Input $/M
    $0.500–$30.00
    On global frontier
    0
    2 free / batch variantsSource notes available
  • OpenAIlineage

    OpenAI gpt-oss family

    OpenAI’s gpt-oss checkpoints are separate downloadable models. The weights, serving runtime and hosted endpoint each bring their own requirements.

    Standard rows
    3
    Rated
    2 / 3
    Input $/M
    $0.018–$0.075
    On global frontier
    2
    2 free / batch variantsSource notes available
  • OpenAIaudio

    OpenAI GPT Audio family

    Audio-input and audio-output API products have different billing dimensions from ordinary text requests. A text-token estimate alone is incomplete.

    Standard rows
    2
    Rated
    0 / 2
    Input $/M
    $0.600–$2.50
    On global frontier
    0
    0 free / batch variantsSource notes available
  • Anthropiclineage

    Anthropic Claude family

    Opus, Sonnet, Haiku and Fable are distinct named rows. Direct API retirement dates, partner-platform schedules and restricted model access require separate checks.

    Standard rows
    15
    Rated
    15 / 15
    Input $/M
    $1.00–$15.00
    On global frontier
    1
    14 free / batch variantsSource notes available
  • Qwenportfolio

    Alibaba Qwen family

    Qwen covers hosted services, downloadable checkpoints, coding and multimodal models. Instruct, Thinking and dated revisions retain separate identities and rating evidence.

    Standard rows
    53
    Rated
    22 / 53
    Input $/M
    $0.030–$4.00
    On global frontier
    0
    0 free / batch variantsSource notes available
  • DeepSeekportfolio

    DeepSeek family

    DeepSeek checkpoint names and hosted aliases can diverge. Current direct-API documents conflict about the Pro alias; the catalogue rows below retain their accepted identities.

    Standard rows
    14
    Rated
    11 / 14
    Input $/M
    $0.053–$1.30
    On global frontier
    2
    1 free / batch variantsSource notes available
  • Metaportfolio

    Meta Llama family

    Meta chat checkpoints, moderation models and third-party fine-tunes are different products. Llama licence terms and endpoint capabilities need a model-specific check.

    Standard rows
    8
    Rated
    5 / 8
    Input $/M
    $0.020–$0.720
    On global frontier
    1
    0 free / batch variantsSource notes available
  • Mistralportfolio

    Mistral, Codestral, Devstral and Voxtral family

    Generalist, coding, audio and specialist Mistral models are grouped by their actual task. A provider hosting a model is not necessarily the model author.

    Standard rows
    15
    Rated
    3 / 15
    Input $/M
    $0.018–$2.00
    On global frontier
    0
    5 free / batch variantsSource notes available
  • Moonshot AIlineage

    Moonshot Kimi family

    Kimi generations and Thinking-labelled rows are distinct. Check the exact model’s reasoning, tool-message and licence requirements before migrating an integration.

    Standard rows
    7
    Rated
    3 / 7
    Input $/M
    $0.465–$0.680
    On global frontier
    0
    1 free / batch variantsSource notes available
  • xAIlineage

    xAI Grok family

    Grok generations and Fast-labelled rows retain their own evidence. Reasoning effort and multi-agent behaviour are model-specific controls, not a shared family setting.

    Standard rows
    7
    Rated
    4 / 7
    Input $/M
    $1.00–$2.00
    On global frontier
    0
    1 free / batch variantsSource notes available
  • Coherelineage

    Cohere Command family

    Cohere Command generations and dated revisions are separate rows. Embed, Rerank and North products are not interchangeable Command chat models.

    Standard rows
    5
    Rated
    2 / 5
    Input $/M
    $0.038–$2.50
    On global frontier
    0
    0 free / batch variantsSource notes available
  • Googlelineage

    Google Gemma family

    Gemma generation, size and checkpoint labels describe different architectures. Open-weight status alone does not establish a shared licence or identical audio support.

    Standard rows
    6
    Rated
    6 / 6
    Input $/M
    $0.050–$0.650
    On global frontier
    2
    2 free / batch variantsSource notes available
  • Amazonlineage

    Amazon Nova family

    Nova Micro, Lite, Pro and Premier have different documented roles and input support. Vendor positioning is context for a trial, not an independent quality ranking.

    Standard rows
    5
    Rated
    1 / 5
    Input $/M
    $0.035–$2.50
    On global frontier
    0
    0 free / batch variantsSource notes available
  • Perplexityproduct line

    Perplexity Sonar family

    Sonar search-grounded answers, reasoning products and Deep Research have distinct request behaviour. Raw search and third-party Agent API offerings are separate services.

    Standard rows
    5
    Rated
    0 / 5
    Input $/M
    $1.00–$3.00
    On global frontier
    0
    0 free / batch variantsSource notes available
  • Googlemusic

    Google Lyria family

    These are music-generation products. Token prices alone cannot establish a comparable cost per finished audio result or a substitute for a text model.

    Standard rows
    2
    Rated
    0 / 2
    Input $/M
    Unknown
    On global frontier
    0
    0 free / batch variantsSource notes available
  • Xiaomilineage

    Xiaomi MiMo collection

    These MiMo rows are present in the accepted catalogue. The reviewed official source did not establish every current marketplace API identity; exact endpoint confirmation remains necessary.

    Standard rows
    5
    Rated
    4 / 5
    Input $/M
    $0.140–$4.35
    On global frontier
    2
    0 free / batch variantsSource notes available
  • Metaportfolio

    Meta Muse collection

    Muse and Llama are separate Meta product lines. Contributor-labelled access must retain its own service and data-use terms; a lower listed price does not establish equivalence.

    Standard rows
    6
    Rated
    3 / 6
    Input $/M
    $0.100–$1.25
    On global frontier
    0
    0 free / batch variantsSource notes available
  • Tencentportfolio

    Tencent Hunyuan collection

    Tencent’s Hunyuan portfolio includes general and translation-labelled checkpoints. Compare the named task and accepted endpoint capabilities before treating them as substitutes.

    Standard rows
    7
    Rated
    1 / 7
    Input $/M
    $0.044–$0.834
    On global frontier
    0
    0 free / batch variantsSource notes available
  • ByteDancelineage

    ByteDance Seed collection

    Seed Pro, Lite and related rows retain their own checkpoint and endpoint evidence. A vendor’s upgraded model description does not automatically update every reseller endpoint.

    Standard rows
    6
    Rated
    0 / 6
    Input $/M
    $0.075–$0.500
    On global frontier
    0
    0 free / batch variantsSource notes available
  • IBMlineage

    IBM Granite collection

    Compare the accepted IBM Granite catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    1 / 2
    Input $/M
    $0.017–$0.060
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Inceptionlineage

    Inception Mercury collection

    Compare the accepted Inception Mercury catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    1 / 2
    Input $/M
    $0.200–$0.250
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Upstagelineage

    Upstage Solar collection

    Compare the accepted Upstage Solar catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    3
    Rated
    1 / 3
    Input $/M
    $0.100–$0.300
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Rekaportfolio

    Reka collection

    Compare the accepted Reka catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    0 / 2
    Input $/M
    $0.100–$0.100
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • inclusionAIlineage

    InclusionAI Ling collection

    Compare the accepted InclusionAI Ling catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    3
    Rated
    0 / 3
    Input $/M
    $0.060–$0.075
    On global frontier
    0
    1 free / batch variantsCatalogue evidence only
  • Nous Researchlineage

    Nous Hermes collection

    Compare the accepted Nous Hermes catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    3
    Rated
    0 / 3
    Input $/M
    $0.700–$1.00
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Morphlineage

    Morph collection

    Compare the accepted Morph catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    0 / 2
    Input $/M
    $0.800–$0.900
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Poolsidelineage

    Poolside Laguna collection

    Compare the accepted Poolside Laguna catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    0 / 2
    Input $/M
    $0.100–$0.100
    On global frontier
    0
    2 free / batch variantsCatalogue evidence only
  • Thinking Machineslineage

    Thinking Machines Inkling collection

    Compare the accepted Thinking Machines Inkling catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    2 / 2
    Input $/M
    $0.450–$0.950
    On global frontier
    0
    2 free / batch variantsCatalogue evidence only
  • Kwaipilotlineage

    KwaiPilot KAT Coder collection

    Compare the accepted KwaiPilot KAT Coder catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    1
    Rated
    0 / 1
    Input $/M
    $0.740–$0.740
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Inference Netlineage

    Inference.net Schematron collection

    Compare the accepted Inference.net Schematron catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    0 / 2
    Input $/M
    $0.030–$0.050
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Sakana AIlineage

    Sakana Fugu collection

    Compare the accepted Sakana Fugu catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    3
    Rated
    0 / 3
    Input $/M
    $2.00–$5.00
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Aion Labslineage

    Aion collection

    Compare the accepted Aion catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    5
    Rated
    0 / 5
    Input $/M
    $0.700–$3.00
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only
  • Sao10klineage

    Sao10K Euryale collection

    Compare the accepted Sao10K Euryale catalogue rows by exact checkpoint, capability and price. This collection organises related names; official equivalence, shared licences and shared benchmark results are not established.

    Standard rows
    2
    Rated
    0 / 2
    Input $/M
    $0.650–$0.850
    On global frontier
    0
    0 free / batch variantsCatalogue evidence only

What this index covers

Membership follows exact accepted catalogue IDs. Official notes support only the named checkpoint or platform, not every reseller in a collection. The price range uses known listed input rates before context tiers, cached tokens and other billing dimensions. A family can contain models for different jobs; its range is not a like-for-like saving.

Unrated models stay visible. A zero frontier count is a result for the global balanced view, not a verdict on your task.

40 standard rows outside these collections

These rows remain in the full catalogue. Being ungrouped does not mean unavailable or lower quality.

Evidence & Ask
Markdown for LLMs