Models

AI models

Canonical model releases across every lab and modality. Parameters, context, openness and licences exactly as their sources state them — one row per real model, artifacts folded under their canonical release.

339 canonical models (artifacts excluded)

Understood asreleased ≥ 2022 identity trust = Identity probable clear all

Sort: Recently updated · Release date · Name · Parameters · Context window · Cheapest output · Data quality · Downloads

201–250 of 339 · canonical models

Models
#	Model	Params	Context	Openness	Licence	Released	Best price in / out	Data quality
201	DeepSeek-R1-0528-Qwen3-8BDeprecatedIdentity probable DeepSeek Qwen3	8.19B	32.8K	Open weights	MIT	29 May 2025	—	63
202	DeepSeek-R1-Distill-Qwen-32BDeprecatedIdentity probable DeepSeek Qwen	32.8B	128K	Open weights	MIT	20 Jan 2025	—	63
203	DeepSeek-R1-Distill-Qwen-1.5BDeprecatedIdentity probable DeepSeek Qwen	1.78B	128K	Open weights	MIT	20 Jan 2025	—	63
204	granite-4.1-30bIdentity probable IBM Granite 4.1	28.9B	131.1K	Open weights	Apache-2.0	6 Apr 2026	—	63
205	granite-4.1-3bIdentity probable IBM Granite 4.1	3.4B	131.1K	Open weights	Apache-2.0	6 Apr 2026	—	63
206	granite-4.1-8bIdentity probable IBM Granite 4.1	8.79B	131.1K	Open weights	Apache-2.0	6 Apr 2026	—	63
207	LFM2.5-1.2B-InstructIdentity probable Liquid AI LFM2.5	1.17B	32K	Open weights	Other	6 Jan 2026	—	63
208	LFM2.5-VL-1.6BIdentity probable Liquid AI LFM2.5	1.6B	32K	Open weights	Other	5 Jan 2026	—	63
209	LFM2.5-8B-A1BIdentity probable Liquid AI LFM2.5	8.47B1B active	32.8K	Open weights	Other	28 May 2026	—	63
210	LFM2-1.2BDeprecatedIdentity probable Liquid AI LFM2.1	1.17B	32.8K	Open weights	Other	10 Jul 2025	—	63
211	Llama-3.2-1BDeprecatedIdentity probable Meta AI Llama 3.2	1.24B	128K	Restricted weights	Llama-3.2-Community	18 Sept 2024	—	63
212	Llama-3.2-3BDeprecatedIdentity probable Meta AI Llama 3.2	3.21B	128K	Restricted weights	Llama-3.2-Community	18 Sept 2024	—	63
213	Llama-3.1-405BIdentity probable Meta AI Llama 3.1	405.9B	128K	Restricted weights	Llama-3.1-Community	16 Jul 2024	—	63
214	MiniMax-M1-40kDeprecatedIdentity probable MiniMax MiniMax	456.1B	1M	Open weights	Apache-2.0	5 Jun 2025	—	63
215	MiniMax-M1-80kDeprecatedIdentity probable MiniMax MiniMax	456.1B	1M	Open weights	Apache-2.0	13 Jun 2025	—	63
216	Kimi-Linear-48B-A3B-InstructIdentity probable Moonshot AI Kimi	49.1B3B active	1M	Open weights	MIT	30 Oct 2025	—	63
217	Qwen3.5-4BIdentity probable Qwen Qwen3.5	4.66B	262.1K	Open weights	Apache-2.0	27 Feb 2026	—	63
218	Qwen3-VL-4B-InstructDeprecatedIdentity probable Qwen Qwen3	4.44B	256K	Open weights	Apache-2.0	11 Oct 2025	—	63
219	grok-2DeprecatedIdentity probable xAI Grok 2	—	131.1K	Open weights	—	22 Aug 2025	—	57
220	grok-1DeprecatedIdentity probable xAI Grok 1	—	8.19K	Open weights	Apache-2.0	17 Mar 2024	—	57
221	QwQ-32BAnnouncedIdentity probable Alibaba Group Qwen	32B	131.1K	Open weights	Apache-2.0	6 Mar 2025	—	68
222	qwen-2-5-maxIdentity probable Alibaba Group Qwen2.5	—	32K	Closed	—	28 Jan 2025	—	63
223	QwQ-32B-PreviewPreviewIdentity probable Alibaba Group Qwen	32B	32.8K	Open weights	—	28 Nov 2024	—	60
224	deepseek-v4-proIdentity probable DeepSeek DeepSeek	—	1M	Closed	—	16 Aug 2026	—	60
225	qwen-turboIdentity probable Alibaba Group Qwen	—	1M	Closed	—	15 Nov 2024	—	63
226	qwen3-coder-480b-a35b-instructAnnouncedIdentity probable Alibaba Group Qwen3	480B35B active	256K	Open weights	—	22 Jul 2025	—	66
227	WhisperDeprecatedIdentity probable OpenAI Whisper	1.54B	—	Closed	Apache-2.0	7 Nov 2023	—	72
228	DeepSeek V4 Flash (0731)Identity probable DeepSeek DeepSeek	304.2B	1.31M	Open weights	MIT	31 Jul 2026	—	72
229	DeepSeek V4 Flash Vision ExpPreviewIdentity probable DeepSeek DeepSeek	304.6B	1.05M	Open weights	MIT	31 Aug 2026	—	72
230	Hy-MT2-1.8BIdentity probable Tencent	2.04B	8.19K	Open weights	Apache-2.0	11 May 2026	—	68
231	Hy-MT2-30B-A3BIdentity probable Tencent	30.1B3B active	8.19K	Open weights	Apache-2.0	11 May 2026	—	68
232	Hy-MT2-7BIdentity probable Tencent	8.03B	8.19K	Open weights	Apache-2.0	11 May 2026	—	68
233	DeepSeek V4 Pro 0423Identity probable DeepSeek DeepSeek	1.6T	1.05M	Open weights	MIT	22 Apr 2026	—	68
234	DeepSeek V4 Flash 0423Identity probable DeepSeek DeepSeek	290.9B	1.05M	Open weights	MIT	22 Apr 2026	—	68
235	Hy3 previewDeprecatedIdentity probable Tencent	298.8B	262.1K	Open weights	Other	13 Apr 2026	—	72
236	Ministral 3 14B 2512Identity probable Mistral AI Ministral 3	13.9B	262.1K	Open weights	Apache-2.0	31 Oct 2025	—	68
237	Ministral 3 3B 2512Identity probable Mistral AI Ministral 3	3.85B	131.1K	Open weights	Apache-2.0	31 Oct 2025	—	68
238	DeepSeek V3.2Identity probable DeepSeek DeepSeek	685.4B	163.8K	Open weights	MIT	1 Dec 2025	—	72
239	gpt-oss-safeguard-20bIdentity probable OpenAI gpt-oss	21.5B	131.1K	Open weights	Apache-2.0	18 Sept 2025	—	74
240	UI-TARS 7BIdentity probable ByteDance	8.29B	128K	Open weights	Apache-2.0	16 Apr 2025	—	68
241	Hunyuan A13B InstructIdentity probable Tencent Hunyuan	80.4B13B active	131.1K	Open weights	Other	25 Jun 2025	—	68
242	ERNIE 4.5 VL 424B A47BIdentity probable Baidu ERNIE 4.5	423.5B47B active	123K	Open weights	Apache-2.0	28 Jun 2025	—	68
243	Mistral Small 3.2 24BIdentity probable Mistral AI Mistral	24B	256K	Open weights	Apache-2.0	19 Jun 2025	—	68
244	Qwen3 8BDeprecatedIdentity probable Qwen Qwen3	8.19B	131.1K	Open weights	Apache-2.0	27 Apr 2025	—	72
245	Qwen3 32BDeprecatedIdentity probable Qwen Qwen3	32.8B	131.1K	Open weights	Apache-2.0	27 Apr 2025	—	72
246	Qwen3 235B A22BIdentity probable Qwen Qwen3	235B22B active	131.1K	Open weights	Apache-2.0	29 Apr 2025	—	71
247	Mistral Small 3.1 24BIdentity probable Mistral AI Mistral	24B	128K	Open weights	Apache-2.0	11 Mar 2025	—	68
248	R1Identity probable DeepSeek	684.5B	64K	Open weights	MIT	20 Jan 2025	—	68
249	Llama 3.3 70B InstructIdentity probable Meta AI Llama 3.3	70.5B	131.1K	Restricted weights	Llama-3.3-Community	26 Nov 2024	—	68
250	Qwen2.5 7B InstructIdentity probable Qwen Qwen2.5	7.62B	32.8K	Open weights	Apache-2.0	16 Sept 2024	—	68

Best price = cheapest current offer across providers when the listing carries one (USD per 1M tokens); otherwise open the model for its provider deployments. Data quality is how well AI Atlas knows the entity, not how good the model is. Rows focus with the keyboard: ↑ ↓ move, Enter inspects. Methodology →