Show me AI models

199 AI models from 51 labs. 78 have weights you can download. Input prices run from $0.03 to $30 per 1M tokens. The largest context window is 2M tokens (Grok 4.20). Newest check: 13 Nov 2026.

87 of 199 models were checked against the vendor's own page in the last 30 days.

ModelLabContextIn / out per 1MReads and writesStatusChecked
GPT-5.5GPT-5.5 is OpenAI's April 2026 flagship model for agentic coding and knowledge work, with a 1.05M-token API context window and text and image input.OpenAI1.1M$5 / $30text, image → textAvailable24 Oct
Qwen3.6-27BQwen3.6-27B, released April 22, 2026 by Alibaba's Qwen Team, is a 27B open-weight dense model with a 262,144-token context window.Alibaba Cloud262K--text, image, video → textAvailable28 Oct
GPT Image 2OpenAI's April 2026 image flagship integrates O-series reasoning for 99% text accuracy, natural language inpainting, and 2K resolution via a token-based API.OpenAI--$8 / $30text, image → imageAvailable30 Aug
Claude Opus 4.7Claude Opus 4.7 by Anthropic (April 2026): 1M-token context, 87.6% SWE-bench Verified, high-res vision, and adaptive thinking at $5/$25 per 1M tokens.Anthropic1M$5 / $25text, image, pdf → textAvailable23 Oct
Gemma 4 31BGemma 4 31B, released April 2, 2026 by Google DeepMind, is a 31B open-weight multimodal model with a 262,144-token context window.Google262K--text, image → textAvailable16 Sep
GPT-5.4GPT-5.4 (OpenAI, March 2026) is a flagship reasoning and coding model with a 1M-token context window and a native Computer Use API.OpenAI1M$2.50 / $15text, image, code → text, codeAvailable27 Aug
Gemini 3.1 Flash-LiteGemini 3.1 Flash-Lite (Google DeepMind, GA May 2026) is the cheapest model in the Gemini 3 lineup, built for high-volume, latency-sensitive tasks like classification, translation, and lightweight agent steps rather than frontier reasoning.Google1M$0.25 / $1.50text, image, audio, video, pdf → textAvailable8 Sep
Gemini 3.1 ProGemini 3.1 Pro tops the GPQA Diamond scores of Claude Opus and GPT-5.2 on Google’s own model card, and takes a million tokens of mixed text, image, audio and video input.Google1M$2 / $12text, image, audio, video, pdf → textPreview23 Oct
Claude Sonnet 4.6Claude Sonnet 4.6 by Anthropic (Feb 2026) scores 79.6% on SWE-bench Verified with a 1M-token context window, and is now one generation behind Sonnet 5.5.Anthropic1M$3 / $15text, image, pdf → textAvailable23 Oct
Grok 4.20Grok 4.20 by xAI (Feb 2026): 4-agent MoE architecture, 2M token context, 88.9% GPQA Diamond. Fastest frontier output at 265 tokens/sec.xAI2M$1.25 / $2.50text, image, pdf → textAvailable27 Aug
Seed 2.0 ProByteDance flagship omni-modal agent (Feb 2026), 272K context, 98.3 AIME 2025, 88.9 GPQA Diamond, 76.5 SWE-bench, native audio and video input.ByteDance272K$0.47 / $2.37text, image, audio, video → textAvailable27 Aug
Seedream 5.0 LiteByteDance Seed's February 2026 Lite model reasons over prompts and can search the web; Artificial Analysis scores it 1020 Elo for text-to-image.ByteDance----text, image → imageAvailable11 Nov
Claude Opus 4.6Claude Opus 4.6, released February 5, 2026, has a 1M-token context window and 128k max output, priced at $5/$25 per 1M tokens and listed Active by Anthropic.Anthropic1M$5 / $25text, image, pdf → textAvailable23 Oct
Riverflow 2.0 ProSourceful's Riverflow 2.0 Pro (February 2026) tops Artificial Analysis's leaderboards for both text-to-image and image editing. Agentic self-correction pipeline, 4K output, and accurate in-image text rendering built for commercial brand design.Sourceful8K--text, image → imageAvailable30 Aug
ALIA-40b-instruct-2601BSC's open instruct model for Spanish, Catalan, Basque and Galician, post-trained with supervised and preference data and released as Apache-licensed weights.Barcelona Supercomputing Center164K--text → textBeta12 Oct
Kanana-2 30B-A3B InstructOpen-weight mixture-of-experts chat model from Kakao, released January 2026, with 3B active parameters and six supported languages.Kakao33K--text → textAvailable13 Oct
HyperCLOVA XNaver's HyperCLOVA X family runs from free SEED weights on Hugging Face to managed CLOVA Studio models, led by an open reasoning model that reads text, images and video.Naver128K--text, image, video → textAvailable31 Oct
Nemotron 3 Nano 30B A3BNVIDIA's open Nemotron 3 Nano (30B total, 3.6B active) hits 78.3% MMLU-Pro with 3.3x the throughput of Qwen3-30B-A3B, released December 2025.NVIDIA1M$0.05 / $0.20text → textAvailable31 Aug
Olmo 3.1 32B ThinkDense 32.2B reasoning model from Ai2 that ships open weights alongside its training data and code, with a 65,536-token window.Ai266K--text → textAvailable3 Oct
Ministral 3 8BMinistral 3 8B by Mistral AI (Dec 2025): Apache 2.0, 8B params, 256K context, 70.6% MMLU-Pro, 78.7% AIME 2025 (reasoning), vision input, $0.15/1M tokens.Mistral AI256K$0.15 / $0.15text, image → textAvailable30 Aug
Seedream 4.5ByteDance's Seedream 4.5, announced by BytePlus on 3 December 2025, takes up to 10 reference images; Artificial Analysis scores it 1030 Elo for text-to-image.ByteDance----text, image → imageAvailable11 Nov
Amazon Nova 2 ProAmazon Nova 2 Pro (Dec 2025) is AWS's most intelligent Nova 2 model, built for multi-step reasoning and agentic tasks with a 1M-token context window.Amazon1M$1.25 / $10text, image, pdf, video → textPreview25 Oct
Ministral 3 14BMinistral 3 14B is Mistral AI's flagship small open-weight model, with a dedicated reasoning variant built for math and science benchmarks.Mistral AI256K$0.20 / $0.20text, image → textAvailable30 Oct
Ministral 3 3BMinistral 3 3B by Mistral AI (Dec 2025): Apache open-weight edge model, 3.4B params, vision input, fits in 8GB VRAM.Mistral AI256K$0.10 / $0.10text, image → textAvailable30 Aug
Mistral Large 3Mistral Large 3 (Dec 2025) is a 675B MoE model (41B active) with 256K context and Apache 2.0 open weights.Mistral AI256K$0.50 / $1.50text, image → textAvailable26 Sep
FLUX.2FLUX.2 by Black Forest Labs (Nov 2025) is a 32B rectified flow transformer family producing up to 4MP images. Five variants span open-weight klein to grounded [max], with multi-reference editing and hex color control.Black Forest Labs----text, image → imageAvailable4 Oct
Gemini 3 Pro ImageNano Banana Pro (Gemini 3 Pro Image), released Nov 20, 2025 by Google DeepMind, generates 4K images with 94% text accuracy and Search-grounded facts.Google66K$2 / $12text, image → image, textAvailable1 Sep
Veo 3.1Google DeepMind's Veo 3.1 generates 4, 6 or 8-second video clips with native synchronized dialogue and audio, rendering up to 4K.Google----text, image → video, audioPreview26 Sep
Osmosis-Structure-0.6BOpen-weight 0.6B model built on Qwen3 that extracts schema-valid structured output from reasoning traces, with 996 monthly Hugging Face downloads.Osmosis----text → textAvailable15 Sep
Phi-4Phi-4 is Microsoft's 14B dense model under the MIT license, scoring 84.8% on MMLU in the vendor's own table. It runs locally through Ollama or as a serverless Azure Foundry endpoint.Microsoft16K$0.125 / $0.50text → textAvailable29 Sep
Recraft V3Recraft's former flagship image model, a previous generation since V4, still the API line for inpainting and exact text placement at 1K output.Recraft----text, image → imageAvailable12 Nov
Gemini NanoGemini Nano is Google's on-device model, now at nano-v4 (Aug 2026); Google published sizes only for its 2023 release: 1.8B and 3.25B parameters.Google4KFreetext, image, audio → textAvailable30 Oct
Mercury 2.5Mercury 2.5 is Inception's diffusion LLM, released 8 September 2026, with four reasoning levels, parallel tool calls and text-only input via API.Inception260K$0.20 / $0.75text → textDeprecated3 Oct
Gemini 3.7 FlashGoogle DeepMind's August 2026 Flash model for coding and agents with a 1M-token context window; Google now recommends a newer Flash model and routes requests to it.Google1M$0.75 / $3.75text, image, audio, video → textDeprecated26 Oct
Nemotron 3.5 LightningNVIDIA's open MoE model for AI agents: 3B of 30B parameters active per token, a 1M-token window and quicker agent tasks than a same-size Qwen rival on PinchBench.NVIDIA1M$0.025 / $0.12text, code → textAvailable6 Nov
DeepSeek-V4-Flash-0731DeepSeek-V4-Flash-0731 is a retired DeepSeek MoE model with a 1M-token context; its old API name now routes to the successor model.DeepSeek1M$0.15 / $0.60text → textRetired26 Oct
FLUX 3FLUX 3 is a video, image, and audio model from Black Forest Labs, launched July 2026, built on a shared Self-Flow architecture.Black Forest Labs----text, image, video → video, audioPreview18 Oct
Kimi K32.8T-parameter open-weight MoE model from Moonshot AI (July 2026) with a 1M-token context window and the top published open-weight result on GPQA Diamond.Moonshot AI1M$3 / $15text, image → text, codeAvailable23 Sep
DeepSeek-V4 FlashRetired by DeepSeek on 10 September 2026; replaced by V4.1 Flash. Legacy API name now routes to the new model. MIT weights remain on Hugging Face.DeepSeek1M$0.15 / $0.60text → textRetired9 Nov
MiniMax M2.7MiniMax M2.7 is a 230B MoE model (10B active) from March 2026 with a 204,800-token context and 56.2% on SWE-Bench Pro, priced at $0.30/$1.20 per 1M tokens.MiniMax205K$0.30 / $1.20text, code → text, codeAvailable28 Sep
Kling 3.0Kling 3.0 is Kuaishou Technology's proprietary AI video model, built for native 4K output with synchronized multilingual audio in one generation pass.Kuaishou Technology----text, image, video → video, audioAvailable30 Aug
GPT-5.2OpenAI's fourth GPT-5 variant: 400,000-token context, 80% SWE-bench Verified, 92.4% GPQA Diamond. Released December 2025, deprecated June 2026.OpenAI400K$1.75 / $14text, image, audio, video, pdf → textDeprecated27 Aug
Grok 4.1xAI's Nov 17, 2025 upgrade to Grok 4: #1 on LMArena at launch, a sharp hallucination cut, and since deprecated.xAI256K$3 / $15text, image → textDeprecated7 Oct
Grok 4 FastGrok 4 Fast was xAI's cost-optimized 2 million token context model from 2025, since taken off the xAI API in favor of Grok 4.3.xAI2M$0.20 / $0.50text, image → textRetired23 Oct
GPT-5OpenAI's GPT-5 is a Mixture-of-Experts flagship model that scored 74.9% SWE-bench Verified at launch, with a 272K-token context window.OpenAI272K$0.625 / $5text, image, audio, video, pdf → text, audioDeprecated27 Aug
Mistral Medium 3Mistral Medium 3 (May 2025) is a multimodal model scoring 92.1% on HumanEval with a 128K-token context window.Mistral AI128K$0.40 / $2text, image → textDeprecated27 Aug
GPT-4o miniOpenAI's smallest 2024-era multimodal model, built for high-volume, low-cost vision and tool-calling workloads rather than frontier reasoning.OpenAI128K$0.15 / $0.60text, image → textDeprecated27 Aug
GPT-4oGPT-4o is OpenAI's 2024 multimodal model with 90.2% HumanEval scoring and real-time voice, now deprecated in favor of newer OpenAI and rival models.OpenAI128K$2.50 / $10text, image, audio, video → text, audio, imageDeprecated27 Aug
Claude 1.0Claude 1.0 was Anthropic's first widely offered model: text-only, trained with Constitutional AI, and since retired. This page records what it was and what replaced it.Anthropic9K--text → textRetired14 Oct

Lab

Licence

A person checks every record against the vendor's own page. Rankings cannot be bought. How Hokai checks records. Not sure what you need? Describe the job to Smart Match and get three picks with reasons.