# AI model releases in March 2026

14 confirmed releases from 11 providers.

| Confirmed | Provider | Release | Models | Evidence |
|---|---|---|---|---|
| Mar 31, 2026 | H Company | Holo3 | [Holo3-122B-A10B](https://benchlm.ai/models/holo3-122b-a10b), [Holo3-35B-A3B](https://benchlm.ai/models/holo3-35b-a3b) | [H Company](https://hcompany.ai/holo3) |
| Mar 31, 2026 | LiquidAI | LFM2.5-350M | [LFM2.5-350M](https://benchlm.ai/models/lfm2-5-350m) | [Liquid AI](https://www.liquid.ai/blog/lfm2-5-350m-no-size-left-behind) |
| Mar 31, 2026 | Prism ML | 1-bit Bonsai | [1-bit Bonsai 8B](https://benchlm.ai/models/bonsai-8b), [1-bit Bonsai 4B](https://benchlm.ai/models/bonsai-4b), [1-bit Bonsai 1.7B](https://benchlm.ai/models/bonsai-1-7b) | [Prism ML](https://prismml.com/news/bonsai-8b) |
| Mar 30, 2026 | Inception | Mercury Edit 2 | [Mercury Edit 2](https://benchlm.ai/models/mercury-edit-2) | [Inception](https://www.inceptionlabs.ai/blog/introducing-mercury-edit-2) |
| Mar 19, 2026 | Cursor | Composer 2 | [Composer 2](https://benchlm.ai/models/composer-2) | [Cursor](https://cursor.com/blog/composer-2) |
| Mar 18, 2026 | Xiaomi | MiMo-V2-Pro and MiMo-V2-Omni | [MiMo-V2-Pro](https://benchlm.ai/models/mimo-v2-pro), [MiMo-V2-Omni](https://benchlm.ai/models/mimo-v2-omni) | [Xiaomi](https://mimo.xiaomi.com/mimo-v2-pro) |
| Mar 17, 2026 | OpenAI | GPT-5.4 mini and nano | [GPT-5.4 mini](https://benchlm.ai/models/gpt-5-4-mini), [GPT-5.4 nano](https://benchlm.ai/models/gpt-5-4-nano) | [OpenAI](https://openai.com/index/introducing-gpt-5-4-mini-and-nano/) |
| Mar 10, 2026 | xAI | Grok 4.20 family | [Grok 4.20](https://benchlm.ai/models/grok-4-20-beta), [Grok 4.20 Multi-agent](https://benchlm.ai/models/grok-4-20-multi-agent-beta) | [xAI](https://docs.x.ai/developers/release-notes) |
| Mar 6, 2026 | Sarvam | Sarvam 30B and 105B | [Sarvam 30B](https://benchlm.ai/models/sarvam-30b), [Sarvam 105B](https://benchlm.ai/models/sarvam-105b) | [Sarvam AI](https://www.sarvam.ai/blogs/sarvam-30b-105b) |
| Mar 5, 2026 | OpenAI | GPT-5.4 and GPT-5.4 Pro | [GPT-5.4](https://benchlm.ai/models/gpt-5-4), [GPT-5.4 Pro](https://benchlm.ai/models/gpt-5-4-pro) | [OpenAI](https://openai.com/index/introducing-gpt-5-4/) |
| Mar 4, 2026 | Alibaba | Qwen3.5 122B, 27B, and 35B-A3B | [Qwen3.5-122B-A10B](https://benchlm.ai/models/qwen3-5-122b-a10b), [Qwen3.5-27B](https://benchlm.ai/models/qwen3-5-27b), [Qwen3.5-35B-A3B](https://benchlm.ai/models/qwen3-5-35b-a3b) | [Alibaba Cloud Model Studio](https://www.alibabacloud.com/help/en/model-studio/newly-released-models) |
| Mar 4, 2026 | Alibaba | Qwen3.5 Plus and Flash | [Qwen3.5 Plus](https://benchlm.ai/models/qwen3-5-plus), [Qwen3.5 Flash](https://benchlm.ai/models/qwen3-5-flash) | [Alibaba Cloud Model Studio](https://www.alibabacloud.com/help/en/model-studio/newly-released-models) |
| Mar 3, 2026 | Google | Gemini 3.1 Flash-Lite | [Gemini 3.1 Flash-Lite](https://benchlm.ai/models/gemini-3-1-flash-lite) | [Google](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-flash-lite/) |
| Mar 3, 2026 | OpenAI | GPT-5.3 Instant | [GPT-5.3 Instant](https://benchlm.ai/models/gpt-5-3-instant) | [OpenAI](https://openai.com/index/gpt-5-3-instant/) |

Canonical page: https://benchlm.ai/model-updates/releases/march-2026