AxForge
Sign in Get started

Model catalogue

Explore models

Model families from the labs, their builds and hardware requirements.

Model type

Filters apply together. Press Enter in search or choose Apply filters.

Reset filters GPU catalogue · AxForge model offerings

1095 matching of 3295 lab model families · Showing 1 to 50

Page 1 of 22

Model reference catalogue
Model TaskSizeContext LicenceEst. GPU Released
Qwen3-untied PrimeIntellect · 1 build Qwen3-0.6B-untied Text gen752M 40,960 tokens Apache 2.0 ~3.1 GiB2026-10-05
DiarizationLM-Gemma-4 Google · 1 build DiarizationLM-Gemma-4-E4B-v1 Vision + text8.0B 131,072 tokens Apache 2.0 ~6.7 GiB2026-10-04
Qwen3.5-Reverse-Text-RL PrimeIntellect · 1 build Qwen3.5-0.8B-Reverse-Text-RL Vision + text1.1B 262,144 tokens Apache 2.0 ~4.4 GiB2026-10-04
Qwen3.5-Reverse-Text-SFT PrimeIntellect · 1 build Qwen3.5-0.8B-Reverse-Text-SFT Vision + text1.1B 262,144 tokens Apache 2.0 ~4.4 GiB2026-10-04
Kolibri-1 Aleph-Alpha · 1 build Kolibri-1 Text gen78.1B 262,144 tokens Apache 2.0 ~81.4 GiB2026-10-02
HunyuanImage-3-verbatim-flashpack fal · 1 build HunyuanImage-3-Instruct-verbatim-flashpack Image edit 22,800 tokens Licence fee 2026-10-01
PixelUMM NVIDIA · 1 build PixelUMM Image→text Licence fee 2026-10-01
kitten-tts-2 KittenML · 1 build kitten-tts-2 Text→speech Licence fee ~2.7 GiB2026-09-30
Tokle techdotus · 1 build Tokle-3M Text gen3M 512 tokens MIT ~0.5 GiB2026-09-30
whistle Cactus-Compute · 1 build whistle Speech→text Apache 2.0 ~1.9 GiB2026-09-30
AdvancedMathBench-AutoVerifier internlm · 1 build AdvancedMathBench-AutoVerifier Vision + text36.0B 262,144 tokens ~76.7 GiB2026-09-29
AREX-2 BAAI · 1 build AREX-2 Vision + text27.4B 262,144 tokens Apache 2.0 ~58.5 GiB2026-09-29
Florence-2-large-FlashPack fal · 1 build Florence-2-large-FlashPack Vision + text 4,096 tokens MIT 2026-09-29
ForgePlex-M2 ForgeWorks · 1 build ForgePlex-M2-9M Text gen10M 1,024 tokens Apache 2.0 ~0.5 GiB2026-09-29
ovie-512 kyutai · 1 build ovie-512 Image edit145M MIT ~5.9 GiB2026-09-29
ovie-ft-dl3dv kyutai · 1 build ovie-ft-dl3dv Image edit143M MIT ~5.9 GiB2026-09-29
ovie-ft-re10k kyutai · 1 build ovie-ft-re10k Image edit143M MIT ~5.9 GiB2026-09-29
Tokle-SPAB techdotus · 1 build Tokle-SPAB-3M Text gen3M 512 tokens MIT ~0.5 GiB2026-09-29
IQuest-Q1 IQuestLab · 1 build IQuest-Q1 Text gen320.3B 524,288 tokens Licence fee ~652.2 GiB2026-09-28
Darwin-RSI FINAL-Bench · 2 builds Darwin-27B-RSI Text gen26.9B 262,144 tokens Apache 2.0 ~56.2 GiB2026-09-27
MiMo-Flash-MOPD XiaomiMiMo · 1 build MiMo-V2.6-Flash-MOPD Text gen310.8B 1,048,576 tokens MIT ~179.4 GiB2026-09-27
MiMo-Pro-MOPD XiaomiMiMo · 1 build MiMo-V2.6-Pro-MOPD Text gen1024.2B 1,048,576 tokens MIT ~580.1 GiB2026-09-27
Naive-N0.5-Flash NaiveAI · 1 build Naive-N0.5-Flash Text gen 1,048,576 tokens MIT ~628.5 GiB2026-09-27
Qwen-Image-2.1-Components-FlashPack fal · 1 build Qwen-Image-2.1-Components-FlashPack Text→image Licence fee 2026-09-27
Intern-Decision internlm · 3 builds Intern-Decision-4B Vision + text4.5B 262,144 tokens Apache 2.0 ~11.7 GiB2026-09-26
K2-Type IFM · 1 build K2-Type-0.9B Text gen1.1B 131,072 tokens Apache 2.0 ~3.2 GiB2026-09-26
clio-legacy NovelAI · 1 build clio-v1-legacy Text gen3.0B 8,192 tokens Conditions ~5.1 GiB2026-09-24
Tev1-experimental togethercomputer · 2 builds Tev1-4B-experimental Text gen4.7B 262,144 tokens ~11.9 GiB2026-09-23
ThinkingCap-Qwen3.8 bottlecapai · 1 build ThinkingCap-Qwen3.8-27B-NVFP4 Vision + text16.7B 262,144 tokens Licence fee ~23.9 GiB2026-09-23
tinctura bench-labs · 1 build tinctura-v1 Text gen96M 2,048 tokens Apache 2.0 ~1.9 GiB2026-09-23
parakeet-ultra moondream · 1 build parakeet-ultra Speech→text627M Open ~2.9 GiB2026-09-22
Audio8-ASR-Infinite Edge0 · 1 build Audio8-ASR-Infinite Speech→text4.1B 32,768 tokens Apache 2.0 ~9.9 GiB2026-09-21
limite-violetto paradigma-inc · 1 build limite-1b-violetto Text gen1.0B 131,072 tokens Apache 2.0 ~3.1 GiB2026-09-21
MiMo-V2.6-Flash Xiaomi · 1 build MiMo-V2.6-Flash-RL Text genFlash 1,048,576 tokens MIT ~189.9 GiB2026-09-21
MiMo-Pro-RL XiaomiMiMo · 1 build MiMo-V2.6-Pro-RL Text gen1024.2B 1,048,576 tokens MIT ~580.1 GiB2026-09-21
Ovis-VL-Embedding ATH-MaaS · 1 build Ovis-VL-Embedding-2B Embeddings 262,144 tokens Apache 2.0 ~5.6 GiB2026-09-21
Supra2-IMG SupraLabs · 1 build Supra2-IMG Text→image Apache 2.0 ~6.0 GiB2026-09-21
WorldCrafter TencentARC · 1 build WorldCrafter-Base Image→video14.3B ~30.1 GiB2026-09-21
WorldCrafter-Fast TencentARC · 1 build WorldCrafter-Fast Image→video ~58.4 GiB2026-09-21
Ovis-Omni-Embedding ATH-MaaS · 1 build Ovis-Omni-Embedding-3B Embeddings Apache 2.0 ~12.3 GiB2026-09-20
Qwen-Image-2.1-PE-T2I Qwen · 1 build Qwen-Image-2.1-PE-T2I Text→image9.4B 262,144 tokens Licence fee ~10.8 GiB2026-09-20
Step-5 TypeSafeAI · 1 build Step-5-Preview-BF16 Vision + text604.3B 1,048,576 tokens Licence fee ~1259.3 GiB2026-09-20
tedlium2_streaming_transformer espnet · 1 build tedlium2_streaming_transformer Speech→text Open ~1.8 GiB2026-09-20
LFM2.5-VL-DSpark LiquidAI · 1 build LFM2.5-VL-3B-DSpark Vision + text279M 128,000 tokens Licence fee ~1.3 GiB2026-09-18
parakeet-redux moondream · 1 build parakeet-redux Speech→text149M Open ~1.8 GiB2026-09-18
GAE-D64 TencentARC · 1 build GAE-D64-1B Image→video ~4.3 GiB2026-09-17
GLM-5.3-Flash-Spark local-inference-lab · 1 build GLM-5.3-Flash-NVFP4-Spark Text gen165.5B 1,048,576 tokens MIT ~193.6 GiB2026-09-17
Ming-Image-0.1-Design inclusionAI · 1 build Ming-Image-0.1-Design Text→image6.2B MIT ~40.0 GiB2026-09-17
Vela-1.0-Omni-Mini llm-semantic-router · 1 build Vela-1.0-Omni-Mini Embeddings1.4B Apache 2.0 ~3.2 GiB2026-09-17
Vela-1.0-Omni-Nano llm-semantic-router · 1 build Vela-1.0-Omni-Nano Embeddings164M Apache 2.0 ~1.4 GiB2026-09-17

Release dates use the upstream repository's publication date. The default is the past 12 months, through today. All time includes unknown and future dates. 3 families have no recorded date.

Size, context and GPU memory describe the listed build. For language models, the memory estimate covers one request at 8K context (or the model's shorter limit); image estimates cover one image. Other types use their existing weight and overhead estimates. Blank values are missing or incomplete and do not match numeric limits. Open a model to compare its builds and hardware estimates.

© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms