Reference

Model Directory

Frontier and notable open models — who makes them, when they shipped, and where to try them.

Every model listed here links to its maker's official announcement or documentation — specs come from those sources, and anything we can't verify is left out. For news about these models, see the Models coverage hub. Agents and scripts can pull this directory as JSON.

At a glance

ModelMakerReleasedWeightsLicenseModality
Claude Opus 5 Anthropic 2026-07-24 Closed Multimodal
Claude Sonnet 5 Anthropic 2026-06-30 Closed Multimodal
Claude Fable 5 Anthropic 2026-06-09 Closed Multimodal
Claude Opus 4.8 Anthropic 2026-05-28 Closed Multimodal
Claude Haiku 4.5 Anthropic 2025-10-01 Closed Multimodal
GPT-5.6 OpenAI 2026-07-09 Closed Multimodal
gpt-oss-120b OpenAI 2025-08 Open weights Apache 2.0 Text
gpt-oss-20b OpenAI 2025-08 Open weights Apache 2.0 Text
Gemini 3.6 Flash Google DeepMind 2026-07-21 Closed Multimodal
Gemini 3.5 Flash-Lite Google DeepMind 2026-07-21 Closed Multimodal
Gemma 4 Google DeepMind 2026-04-02 Open weights Apache 2.0 Multimodal
Llama 4 Maverick Meta AI 2025-04-05 Open weights Llama 4 Community License Multimodal
Llama 4 Scout Meta AI 2025-04-05 Open weights Llama 4 Community License Multimodal
DeepSeek-V4-Pro DeepSeek 2026-04 Open weights MIT Multimodal
DeepSeek-V4-Flash DeepSeek 2026-04 Open weights MIT Text
Qwen3.8-Max Alibaba (Qwen) 2026-08-03 Closed Multimodal
Qwen3.6-27B Alibaba (Qwen) 2026-04 Open weights Apache 2.0 Multimodal
Mistral Large 3 Mistral AI 2025-12-02 Open weights Apache 2.0 Multimodal
Grok 4.6 xAI 2026-08-07 Closed Multimodal
MiniMax M2.7 MiniMax 2026-03-18 Open weights Text
Kimi K3 Moonshot AI 2026-07-27 Open weights Kimi K3 License Multimodal

Anthropic

ClosedMultimodal

Claude Opus 5

Anthropic's Opus-tier model for complex agentic coding and enterprise work, positioned near Fable 5's intelligence at roughly half the price. Ships with a 1M-token context window.

Released 2026-07-24
AnnouncementTry itDocs
ClosedMultimodal

Claude Sonnet 5

The balanced tier in the Claude 5 family, tuned for the best combination of speed and intelligence across coding and agent workloads. Carries a 1M-token context window.

Released 2026-06-30
AnnouncementTry itDocs
ClosedMultimodal

Claude Fable 5

Anthropic's most capable widely released model and the first generally available Mythos-class tier, aimed at long-running agents. Priced at $10/$50 per million input/output tokens.

Released 2026-06-09
AnnouncementTry itDocs
ClosedMultimodal

Claude Opus 4.8

The prior Opus flagship, now a legacy model succeeded by Opus 5 but still available via the Claude API. Introduced user-facing effort control and a faster inference mode.

Released 2026-05-28
AnnouncementDocs
ClosedMultimodal

Claude Haiku 4.5

Anthropic's fastest model, offering near-frontier intelligence at low cost with a 200K-token context window. Supports optional extended thinking.

Released 2025-10-01
Try itDocs

OpenAI

ClosedMultimodal

GPT-5.6

OpenAI's current flagship family, shipping in three tiers - Luna, Terra, and Sol - from most cost-efficient to most capable. Sol is positioned as OpenAI's strongest coding and vision model to date.

Released 2026-07-09
Model cardDocs
Open weightsText

gpt-oss-120b

OpenAI's larger open-weight model (about 117B parameters), released under Apache 2.0 for local and self-hosted use with a focus on reasoning tasks.

Released 2025-08 Apache 2.0
Weights
Open weightsText

gpt-oss-20b

OpenAI's smaller open-weight model (about 22B parameters) under Apache 2.0, designed to run efficiently on modest hardware.

Released 2025-08 Apache 2.0
Weights

Google DeepMind

ClosedMultimodal

Gemini 3.6 Flash

Google's high-throughput workhorse model, tuned for lower latency and roughly 17% fewer output tokens than 3.5 Flash. Generally available for production use.

Released 2026-07-21
Docs
ClosedMultimodal

Gemini 3.5 Flash-Lite

The fastest, lowest-cost model in Google's Gemini 3.5 line, aimed at high-volume, latency-sensitive workloads.

Released 2026-07-21
Docs
Open weightsMultimodal

Gemma 4

Google DeepMind's open-weight family (E2B, E4B, 26B MoE, and 31B dense) built from the same research as Gemini 3, now shipped under Apache 2.0. Handles text, images, audio, and video with up to 256K context.

Released 2026-04-02 Apache 2.0
AnnouncementModel cardDocs

Meta AI

Open weightsMultimodal

Llama 4 Maverick

Meta's larger open-weight Llama 4 model, a mixture-of-experts design (about 400B total, 17B active parameters) with native text-and-image input.

Released 2025-04-05 Llama 4 Community License
AnnouncementWeights
Open weightsMultimodal

Llama 4 Scout

The smaller Llama 4 model that fits on a single high-end GPU, with 17B active parameters and a very long context window.

Released 2025-04-05 Llama 4 Community License
AnnouncementWeights

DeepSeek

Open weightsMultimodal

DeepSeek-V4-Pro

DeepSeek's flagship open-weight model, a 1.6T-parameter mixture-of-experts release under the MIT license with image input and a 1M-token context via API.

Released 2026-04 MIT
Weights
Open weightsText

DeepSeek-V4-Flash

The smaller, faster text-only member of the DeepSeek V4 family (about 291B parameters), MIT-licensed and tuned for agentic and coding workloads.

Released 2026-04 MIT
Weights

Alibaba (Qwen)

ClosedMultimodal

Qwen3.8-Max

Alibaba's flagship Qwen model, a 2.4T-parameter mixture-of-experts system (about 95B active) with a 1M-token context and native text-plus-vision input. Currently API-only; Alibaba has announced an open-weights release.

Released 2026-08-03
Announcement
Open weightsMultimodal

Qwen3.6-27B

An open-weight (Apache 2.0) Qwen model aimed at coding, with a vision encoder for multimodal input. Sized at 27B parameters to run on a single high-end GPU.

Released 2026-04 Apache 2.0
Weights

Mistral AI

Open weightsMultimodal

Mistral Large 3

Mistral's open-weight flagship, a 675B-parameter sparse mixture-of-experts model (41B active) under Apache 2.0 with an added vision encoder for image understanding.

Released 2025-12-02 Apache 2.0
Model cardWeights

xAI

ClosedMultimodal

Grok 4.6

xAI's 1.5T-parameter frontier model, a refinement of Grok 4.5 through improved fine-tuning and reinforcement learning rather than a scale increase. Closed-weight, available through the xAI API and Grok apps.

Released 2026-08-07
Try it

MiniMax

Open weightsText

MiniMax M2.7

An open-weight 230B-parameter mixture-of-experts model (about 10B active) built for agentic and coding workflows, positioned as a low-cost frontier option.

Released 2026-03-18
Weights

Moonshot AI

Open weightsMultimodal

Kimi K3

Moonshot AI's open-weight mixture-of-experts model with 2.8T total parameters (104B active) and a 1M-token context, among the largest open models released. Multimodal via a native vision encoder.

Released 2026-07-27 Kimi K3 License
Weights