26 open-weights AI models from 14 labs. Input prices run from $0.05 to $3 per 1M tokens. The largest context window is 1M tokens (Kimi K3.1 Pro). Newest check: 9 Nov 2026.
94 of 197 models were checked against the vendor's own page in the last 30 days.
| Model | Lab | Context | In / out per 1M | Reads and writes | Status | Checked |
|---|---|---|---|---|---|---|
| LFM3Six open-weight models from Liquid AI, each near 2B parameters, spanning text, vision, audio, omni and decision work, launched in October 2026. | Liquid AI | 262K | -- | text, image, video, audio → text, audio | Available | 2 Nov |
| LFM3-1.7BOpen-weight text model from Liquid AI for local agents and tool calling, released 27 October 2026, with a 256K-token window and a 4-bit size near 1 GB. | Liquid AI | 262K | $0.08 / $0.20 | text → text | Available | 4 Nov |
| LFM3-VL-2BOpen-weight vision-language model from Liquid AI that reads images and video on-device, launched 27 October 2026 with a 128K-token context window. | Liquid AI | 128K | $0.10 / $0.50 | text, image, video → text | Available | 4 Nov |
| Kimi K3.1 ProMoonshot AI's October 2026 reasoning model: a 1M-token window, always-on thinking, text and image input, and open weights on Hugging Face. | Moonshot AI | 1M | $2 / $10 | text, image → text | Available | 6 Nov |
| Muse Spark OpenMeta's open-weight model for coding and agents: 816B total parameters, 42B active, released October 2026 under the Muse Model License. | Meta AI | 1M | -- | text, image, video, audio → text | Available | 20 Oct |
| Mistral Large 4Mistral Large 4 is Mistral AI's largest open-weight multimodal MoE model, generally available since November 2026 with weights on Hugging Face. | Mistral AI | 1M | $0.70 / $2.10 | text, image → text | Available | 9 Nov |
| Intern-DecisionShanghai AI Lab's open-weights multimodal decision models (0.8B, 2B, 4B) return typed choices with probabilities in one pass, released Sept 2026. | Shanghai AI Lab | -- | -- | text, image → text | Preview | 18 Oct |
| Qwen-Image-2.1Qwen-Image-2.1 is Alibaba's 7B open-weight diffusion model, released in September 2026 with native transparent image output and multi-image editing. | Alibaba Cloud | -- | -- | text, image → image | Available | 25 Sep |
| Qwen3.8-Flash-NextQwen3.8-Flash-Next is Alibaba's open-weight 125B-parameter MoE model (6B active per token) previewing Qwen4's architecture, with a 262,144-token context window. | Alibaba Cloud | 262K | -- | text, image, video → text | Preview | 27 Aug |
| GLM-5.3GLM-5.3 is Z.ai's text-only flagship for agentic coding and security work, with a 1M-token context, a live API and open weights. | Z.ai | 1M | $1.40 / $4.40 | text → text | Available | 19 Oct |
| LFM2.5-VL-3BLFM2.5-VL-3B is Liquid AI's vision-language model with a 32,768-token context window, released August 2026 to run entirely on-device. | Liquid AI | 33K | -- | text, image → text | Available | 4 Nov |
| Qwen3.8-2.4T-A95BAlibaba's 2.4T-parameter MoE model (95B active), open-weighted August 2026 with a ~1M-token context and a 92.6 GPQA Diamond score. | Alibaba Cloud | 1M | $2 / $6 | text → text | Available | 20 Sep |
| LTX-2.5Lightricks' open-weights video and audio model that renders synchronized 4K HDR clips from text, image or video input, self-hostable and fine-tunable. | Lightricks | -- | -- | text, image, video → video, audio | Available | 15 Oct |
| Qwen3.8-MaxQwen3.8-Max is Alibaba's flagship MoE model, launched Aug 2026, with a 1M-token context, native multimodal input, and a text-only open-weight checkpoint released weeks later. | Alibaba Cloud | 1M | $2 / $6 | text, image, video → text | Available | 14 Sep |
| MiniMax H3MiniMax H3 (Hailuo 3.0) is a 33B open-weight omni-modal model, released July 31 2026, generating 2K/24fps video with native stereo audio in one forward pass. | MiniMax | -- | -- | text, image, video, audio → video, audio | Available | 13 Sep |
| MiniMax Music 3.0MiniMax Music 3.0 is MiniMax's open-weight text-to-music model, released in 2026, generating complete 5-minute songs with performed, not synthesized, vocals. | MiniMax | -- | -- | text → audio | Available | 25 Sep |
| Kimi K2.7 Code HighSpeedOpen-weight 1T-parameter MoE coding model from Moonshot AI (June 2026) with a 256K context window and a HighSpeed mode hitting ~180-260 tokens/sec. | Moonshot AI | 262K | $0.95 / $4 | text, image, video, code → text, code | Beta | 31 Aug |
| Kimi K2.7-CodeOpen-weight 1T/32B-active MoE coding model from Moonshot AI, released June 2026 under a Modified MIT license and built for long-horizon agentic coding. | Moonshot AI | 262K | $0.95 / $4 | text, image, video, code → text, code | Available | 29 Sep |
| MiniMax M3MiniMax's June 2026 open-weight flagship, built around a 1M-token context window, native multimodal input, and frontier-level agentic coding performance. | MiniMax | 1M | $0.30 / $1.20 | text, image, video → text | Available | 29 Sep |
| Kanana-2 30B-A3B InstructOpen-weight mixture-of-experts chat model from Kakao, released January 2026, with 3B active parameters and six supported languages. | Kakao | 33K | -- | text → text | Available | 13 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. | Naver | 128K | -- | text, image, video → text | Available | 31 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. | NVIDIA | 1M | $0.05 / $0.20 | text → text | Available | 31 Aug |
| 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 → image | Available | 4 Oct |
| 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 → text | Available | 15 Sep |
| 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 AI | 1M | $3 / $15 | text, image → text, code | Available | 23 Sep |
| 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. | MiniMax | 205K | $0.30 / $1.20 | text, code → text, code | Available | 28 Sep |
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.