---
title: "AI models released in April 2025"
type: "release-archive"
canonical_url: "https://theaiengineering.com/releases/2025/04"
markdown_url: "https://theaiengineering.com/releases/2025/04.md"
author: "The AI Engineering Editors"
publisher: "The AI Engineering"
updated: "2025-04-30"
citation: "The AI Engineering Editors. \"AI models released in April 2025.\" The AI Engineering, updated Apr 30, 2025. https://theaiengineering.com/releases/2025/04"
license: "CC BY 4.0"
license_url: "https://creativecommons.org/licenses/by/4.0/"
data_csv: "https://theaiengineering.com/releases.csv"
data_json: "https://theaiengineering.com/releases.json"
---

# AI models released in April 2025

40 releases from 21 labs, 19 with open weights. Dates are the labs' official announcement dates.

## April 2025

### Amazon Nova Premier

Amazon Nova Premier is the most capable model in the first Amazon Nova generation, built for complex tasks and to serve as a teacher for distilling smaller custom models, with a one-million-token context window.

- Date: 2025-04-30
- Lab: Amazon
- Type: Multimodal
- Weights: Proprietary
- Modalities: text, image, video → text
- Context window: 1,000,000 tokens
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#amazon-nova-premier

Highlights:

- Accepts text, image and video input; audio is not supported
- One-million-token context window
- First announced as coming on December 3, 2024

Source: [Amazon Nova Premier: Our most capable model for complex tasks and teacher for model distillation](https://aws.amazon.com/blogs/aws/amazon-nova-premier-our-most-capable-model-for-complex-tasks-and-teacher-for-model-distillation/)

### DeepSeek-Prover-V2

DeepSeek-Prover-V2 is an open-weights model for formal theorem proving in Lean 4, trained with cold-start data from a recursive proof-search pipeline powered by DeepSeek-V3. It comes in 7B and 671B sizes.

- Date: 2025-04-30
- Lab: DeepSeek
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Parameters: 7B, 671B
- License: DeepSeek Model License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#deepseek-prover-v2

Highlights:

- DeepSeek-Prover-V2-671B is trained on top of DeepSeek-V3-Base
- DeepSeek-Prover-V2-7B builds on Prover-V1.5-Base with context extended to 32K tokens
- Use is subject to the DeepSeek model license

Source: [DeepSeek-Prover-V2 (GitHub)](https://github.com/deepseek-ai/DeepSeek-Prover-V2)

### Gen-4 References (Gen-4 Image)

Gen-4 image generation model that uses reference images to create consistent characters, locations and other elements, launched as Gen-4 References for paid plans.

- Date: 2025-04-30
- Lab: Runway
- Type: Image
- Weights: Proprietary
- Modalities: text, image → image
- Available at launch: Consumer app, API
- Permalink: https://theaiengineering.com/releases/2025/04#runway-gen-4-references

Highlights:

- Extended to Free plan users on May 7, 2025
- Gen-4 Image API launched May 16, 2025
- Improved object consistency update on June 25, 2025

Source: [Product Updates & Changelog | Runway AI](https://runway.com/changelog)

### MiMo-7B

MiMo-7B is Xiaomi's first open-source language model series, built for math and code reasoning and released as base, SFT, RL-Zero, and RL checkpoints with pre-training and post-training tailored to reasoning.

- Date: 2025-04-30
- Lab: Xiaomi
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Parameters: 7B
- License: MIT
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#mimo-7b

Highlights:

- Base model pre-trained on about 25 trillion tokens with multiple-token prediction
- MiMo-7B-RL-0530 update on May 30, 2025 scaled SFT data and extended the RL window to 48K
- Weights under the MIT license

Source: [XiaomiMiMo/MiMo](https://github.com/XiaomiMiMo/MiMo)

### Phi-4-mini-reasoning

Phi-4-mini-reasoning is a 3.8B open Microsoft model for math reasoning, fine-tuned from Phi-4-mini with synthetic data generated by DeepSeek-R1, and aimed at edge, mobile and educational applications.

- Date: 2025-04-30
- Lab: Microsoft
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: 3.8B
- License: MIT
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2025/04#phi-4-mini-reasoning

Highlights:

- 3.8B parameters with a 128K token context length
- Fine-tuned on synthetic data generated by DeepSeek-R1
- ONNX version for Copilot+ PCs made available May 15, 2025

Source: [One year of Phi: Small language models making big leaps in AI](https://azure.microsoft.com/en-us/blog/one-year-of-phi-small-language-models-making-big-leaps-in-ai/)

### Phi-4-reasoning and Phi-4-reasoning-plus

Phi-4-reasoning is a 14B open-weight reasoning model fine-tuned from Phi-4 on reasoning demonstrations, released with Phi-4-reasoning-plus, which adds reinforcement learning to use more inference-time compute.

- Date: 2025-04-30
- Lab: Microsoft
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Context window: 32,000 tokens
- Parameters: 14B
- License: MIT
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2025/04#phi-4-reasoning

Highlights:

- Both models have 14B parameters and a 32K token context length
- Phi-4-reasoning-plus is further trained with reinforcement learning
- Available on Azure AI Foundry and Hugging Face under the MIT license

Source: [One year of Phi: Small language models making big leaps in AI](https://azure.microsoft.com/en-us/blog/one-year-of-phi-small-language-models-making-big-leaps-in-ai/)

### Llama Guard 4

Llama Guard 4 is a 12B natively multimodal safety classifier that moderates text and multi-image inputs and model responses, acting as a unified safeguard across modalities.

- Date: 2025-04-29
- Lab: Meta
- Type: Multimodal
- Weights: Open weights
- Modalities: text, image → text
- Parameters: 12B
- License: Llama 4 Community License
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2025/04#llama-guard-4

Highlights:

- Dense model pruned from the Llama 4 Scout pretrained model
- Classifies content against 14 MLCommons-based hazard categories
- Also offered on the Llama API, then in limited preview

Source: [Sharing new open source protection tools and advancements in AI privacy and security](https://ai.meta.com/blog/ai-defenders-program-llama-protection-tools/)

### Llama Prompt Guard 2

Llama Prompt Guard 2 updates Meta's prompt attack classifier with an improved 86M model and a smaller 22M model for detecting prompt injection and jailbreak attempts.

- Date: 2025-04-29
- Lab: Meta
- Type: Small model
- Weights: Open weights
- Modalities: text → text
- Context window: 512 tokens
- Parameters: 86M, 22M
- License: Llama 4 Community License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#llama-prompt-guard-2

Highlights:

- The 22M model cuts latency and compute costs by up to 75% versus the 86M model
- 86M is based on mDeBERTa-base; 22M on DeBERTa-xsmall
- Both support a 512-token context window

Source: [Sharing new open source protection tools and advancements in AI privacy and security](https://ai.meta.com/blog/ai-defenders-program-llama-protection-tools/)

### Qwen3

Qwen3 is Alibaba's open-weights model family of two mixture-of-experts models and six dense models from 0.6B to 32B parameters. The models offer hybrid thinking and non-thinking modes with user control over how much reasoning a task receives.

- Date: 2025-04-29
- Lab: Alibaba Qwen
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Context window: 32,768 tokens
- Parameters: MoE: 235B total, 22B active (flagship); 0.6B to 32B dense
- License: Apache-2.0
- Available at launch: API, Consumer app, Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#qwen3

Highlights:

- Pre-trained on about 36T tokens covering 119 languages and dialects
- Flagship Qwen3-235B-A22B supports 32,768 tokens natively and 131,072 with YaRN
- Weights under the Apache 2.0 license

Source: [Qwen3: Think Deeper, Act Faster](https://qwenlm.github.io/blog/qwen3/)

### ERNIE 4.5 Turbo

ERNIE 4.5 Turbo is a faster multimodal model derived from ERNIE 4.5, launched at Baidu Create 2025 with improvements in hallucination reduction, logical reasoning, and coding.

- Date: 2025-04-25
- Lab: Baidu
- Type: Multimodal
- Weights: Proprietary
- Modalities: text, image, video → text
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#ernie-4-5-turbo

Highlights:

- Launched at Baidu Create 2025 in Wuhan
- Enhanced multimodal capabilities, launched alongside ERNIE X1 Turbo
- Serves as the base for the ERNIE X1 Turbo reasoning model

Source: [文心4.5Turbo、X1Turbo和多款AI应用发布！李彦宏：应用创造未来](https://home.baidu.com/home/index/news_detail/id/18023)

### ERNIE X1 Turbo

ERNIE X1 Turbo is a deep-thinking model built on ERNIE 4.5 Turbo, launched at Baidu Create 2025 with a more advanced chain of thought and stronger question answering, writing, reasoning, tool use, and multimodal abilities.

- Date: 2025-04-25
- Lab: Baidu
- Type: Reasoning
- Weights: Proprietary
- Modalities: text, image → text
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#ernie-x1-turbo

Highlights:

- Launched with ERNIE 4.5 Turbo at Create 2025 in Wuhan
- Combines thinking with tool calls in a compound chain of thought and action
- Built on ERNIE 4.5 Turbo, which is derived from ERNIE 4.5

Source: [百度发布文心4.5 Turbo、X1 Turbo两大模型，百度王海峰详解背后技术](https://home.baidu.com/home/index/news_detail/id/18025)

### Kimi-Audio

Kimi-Audio is an open-source audio foundation model for audio understanding, generation, and conversation in one framework. It handles tasks such as speech recognition, audio question answering, captioning, emotion recognition, and end-to-end speech conversation.

- Date: 2025-04-25
- Lab: Moonshot AI
- Type: Speech
- Weights: Open weights
- Modalities: text, audio → text, audio
- Parameters: 7B
- License: MIT
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#kimi-audio

Highlights:

- Pre-trained on more than 13 million hours of audio and text data
- Hybrid input of continuous acoustic and discrete semantic tokens
- Kimi-Audio-7B-Instruct released April 25, 2025; base Kimi-Audio-7B weights followed April 27

Source: [Kimi-Audio (GitHub)](https://github.com/MoonshotAI/Kimi-Audio)

### Step1X-Edit

Step1X-Edit is StepFun's open-source general image editing model that follows natural-language editing instructions, released with inference code, weights, a technical report, and the GEdit-Bench evaluation benchmark.

- Date: 2025-04-25
- Lab: StepFun
- Type: Image
- Weights: Open weights
- Modalities: text, image → image
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#step1x-edit

Highlights:

- Released with the GEdit-Bench image editing benchmark
- Updated v1.1 weights followed on July 9, 2025
- Weights on Hugging Face and ModelScope under Apache-2.0

Source: [GitHub - stepfun-ai/Step1X-Edit](https://github.com/stepfun-ai/Step1X-Edit)

### Lyria 2

Lyria 2 is Google DeepMind's music generation model for high-fidelity, professional-grade audio, introduced with an expanded Music AI Sandbox. It later became generally available to enterprises on Vertex AI and to creators through YouTube Shorts.

- Date: 2025-04-24
- Lab: Google
- Type: Music
- Weights: Proprietary
- Modalities: text → audio
- Available at launch: Consumer app, Cloud platforms
- Permalink: https://theaiengineering.com/releases/2025/04#lyria-2

Highlights:

- All Lyria 2 output is watermarked with SynthID
- Music AI Sandbox access expanded to more U.S. musicians, producers and songwriters
- Generally available on Vertex AI from May 20, 2025

Source: [Music AI Sandbox, now with new features and broader access](https://deepmind.google/discover/blog/music-ai-sandbox-now-with-new-features-and-broader-access/)

### Lyria RealTime

Lyria RealTime is Google DeepMind's interactive music generation model that lets users create, perform and steer a continuous music stream in real time, mixing genres and blending styles as it plays. It powers MusicFX DJ.

- Date: 2025-04-24
- Lab: Google
- Type: Music
- Weights: Proprietary
- Modalities: text → audio
- Available at launch: API, Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#lyria-realtime

Highlights:

- Output is watermarked with SynthID
- Released in the Gemini API and AI Studio as lyria-realtime-exp on May 20, 2025
- Its open-weights counterpart is Magenta RealTime

Source: [Music AI Sandbox, now with new features and broader access](https://deepmind.google/discover/blog/music-ai-sandbox-now-with-new-features-and-broader-access/)

### GLM-ASR

GLM-ASR is Z.ai's speech recognition model on the BigModel API, which uses context to transcribe audio into idiomatic, readable text. Z.ai says it performs better in noisy environments than its earlier models.

- Date: 2025-04-23
- Lab: Z.ai
- Type: Speech
- Weights: Proprietary
- Modalities: audio → text
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#glm-asr

Highlights:

- Supports Chinese, English, and several regional Chinese dialects
- Not disrupted by non-speech noise
- Available through the BigModel API platform

Source: [Model and product release records (BigModel docs)](https://docs.bigmodel.cn/cn/update/new-releases)

### Gemma 3 QAT

Gemma 3 QAT is a set of quantization-aware trained int4 Gemma 3 checkpoints that cut memory needs sharply while preserving quality. The 27B model's weights shrink from 54GB in BF16 to 14.1GB, fitting a single consumer GPU.

- Date: 2025-04-18
- Lab: Google
- Type: LLM
- Weights: Open weights
- Modalities: text, image → text
- Context window: 131,072 tokens
- Parameters: 1B, 4B, 12B, 27B
- License: Gemma Terms of Use
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#gemma-3-qat

Highlights:

- Covers the 27B, 12B, 4B and 1B Gemma 3 sizes
- 27B fits on a single 24GB card such as the NVIDIA RTX 3090
- Official int4 and Q4_0 builds on Hugging Face and Kaggle; supported in Ollama, LM Studio, MLX and llama.cpp

Source: [Gemma 3 QAT Models: Bringing state-of-the-Art AI to consumer GPUs](https://developers.googleblog.com/en/gemma-3-quantized-aware-trained-state-of-the-art-ai-to-consumer-gpus/)

### Gemini 2.5 Flash

Gemini 2.5 Flash is Google's price-performance Gemini 2.5 model, released in preview with a major upgrade in reasoning over 2.0 Flash and a thinking budget that lets developers control how much the model thinks.

- Date: 2025-04-17
- Lab: Google
- Type: LLM
- Weights: Proprietary
- Modalities: text, image, audio, video → text
- Context window: 1,048,576 tokens
- Available at launch: API, Cloud platforms, Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#gemini-2-5-flash

Highlights:

- Preview model gemini-2.5-flash-preview-04-17 in the Gemini API via Google AI Studio and Vertex AI
- An updated preview (gemini-2.5-flash-preview-05-20) shipped at Google I/O 2025
- Stable gemini-2.5-flash became generally available on June 17, 2025
- Also available to everyone in the Gemini app

Source: [Start building with Gemini 2.5 Flash](https://developers.googleblog.com/start-building-with-gemini-25-flash/)

### MAI-DS-R1

MAI-DS-R1 is an open-weights version of DeepSeek R1 post-trained by the Microsoft AI team to respond to more previously blocked topics and reduce risk while keeping the original model's reasoning capabilities.

- Date: 2025-04-17
- Lab: Microsoft
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- License: MIT
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2025/04#mai-ds-r1

Highlights:

- Post-trained from DeepSeek-R1
- Released on Azure AI Foundry and Hugging Face
- Retired from Microsoft Foundry on February 27, 2026

Source: [Introducing MAI-DS-R1](https://techcommunity.microsoft.com/blog/azure-ai-foundry-blog/introducing-mai-ds-r1/4405076)

### Granite 3.3

Update of the Granite 2B and 8B models with improved reasoning that can be toggled, fill-in-the-middle support and a 128K-token context for the 8B model.

- Date: 2025-04-16
- Lab: IBM
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: 2B and 8B
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#granite-3-3

Highlights:

- Released alongside Granite Speech 3.3 8B, an audio-in, text-out speech model
- Base models supersede their Granite 3.1 predecessors
- Released under Apache 2.0

Source: [IBM Granite 3.3: Speech recognition, refined reasoning, and RAG LoRAs](https://www.ibm.com/new/announcements/ibm-granite-3-3-speech-recognition-refined-reasoning-rag-loras)

### OpenAI o3 and o4-mini

o3 and o4-mini are o-series reasoning models and the first that can agentically use and combine every ChatGPT tool, including web search, Python, file analysis, visual reasoning, and image generation.

- Date: 2025-04-16
- Lab: OpenAI
- Type: Reasoning
- Weights: Proprietary
- Modalities: text, image → text
- Context window: 200,000 tokens
- Available at launch: Consumer app, API
- Permalink: https://theaiengineering.com/releases/2025/04#o3

Highlights:

- API model IDs o3 and o4-mini (snapshots dated 2025-04-16)
- Replaced o1, o3-mini, and o3-mini-high for ChatGPT Plus, Pro, and Team users
- ChatGPT also offered o4-mini-high; free users could try o4-mini via 'Think'
- 200,000-token context window per the API model pages

Source: [Introducing OpenAI o3 and o4-mini | OpenAI](https://openai.com/index/introducing-o3-and-o4-mini/)

### Seedream 3.0

Seedream 3.0 is ByteDance Seed's bilingual Chinese-English text-to-image model, adding native 2K output, faster generation, and more accurate small-text rendering and layout compared with Seedream 2.0.

- Date: 2025-04-16
- Lab: ByteDance Seed
- Type: Image
- Weights: Proprietary
- Modalities: text → image
- Available at launch: Consumer app, API
- Permalink: https://theaiengineering.com/releases/2025/04#seedream-3-0

Highlights:

- Native 2K output across aspect ratios without post-processing
- Launched in April 2025 and fully available in Doubao and Jimeng
- Volcano Engine API model ID doubao-seedream-3.0-t2i-250415

Source: [Seedream 3.0 Text-to-Image Model Technical Report Released](https://seed.bytedance.com/en/blog/seedream-3-0-text-to-image-model-technical-report-released)

### UI-TARS-1.5

UI-TARS-1.5 is a multimodal GUI agent built on a vision-language model and improved with reinforcement learning so it reasons before acting in computer-use and game tasks; ByteDance Seed open-sourced a 7B version.

- Date: 2025-04-16
- Lab: ByteDance Seed
- Type: Multimodal
- Weights: Open weights
- Modalities: text, image → text
- Parameters: 7B
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#ui-tars-1-5

Highlights:

- UI-TARS-1.5-7B weights released on Hugging Face under Apache-2.0
- Adds game scenarios supporting up to 1,000 interactions
- The open 7B model targets general computer use rather than games

Source: [GitHub - bytedance/UI-TARS: Pioneering Automated GUI Interaction with Native Agents](https://github.com/bytedance/UI-TARS)

### Embed 4

Multimodal embedding model for searching complex documents such as PDF reports and slides that mix text, images, tables, graphs and code.

- Date: 2025-04-15
- Lab: Cohere
- Type: Embedding
- Weights: Proprietary
- Modalities: text, image → embedding
- Context window: 128,000 tokens
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#cohere-embed-4

Highlights:

- Embeds documents up to 128K tokens, about 200 pages
- Supports 100+ languages and cross-lingual search
- Matryoshka embeddings for flexible storage

Source: [Introducing Embed 4: Multimodal search for business | Cohere Blog](https://cohere.com/blog/embed-4)

### Kling 2.0

Kling 2.0 Master is the second-generation Kling video generation model, launched globally with improvements in motion quality, semantic responsiveness, and visual aesthetics, alongside the Kolors 2.0 image model.

- Date: 2025-04-15
- Lab: Kuaishou
- Type: Video
- Weights: Proprietary
- Modalities: text, image → video
- Available at launch: Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#kling-2-0

Highlights:

- Launched at Kling's From Vision to Screen event in Beijing
- Released together with Kolors 2.0 and new multimodal editing features
- Kuaishou reported more than 20 Kling iterations since June 2024

Source: [Kling AI Advances to the 2.0 Era, Empowering Everyone to Tell Great Stories with AI](https://ir.kuaishou.com/news-releases/news-release-details/kling-ai-advances-20-era-empowering-everyone-tell-great-stories)

### Kolors 2.0

Kolors 2.0 is Kuaishou's second-generation image generation model, launched on Kling AI with the Kling 2.0 video model and improved prompt adherence, cinematic quality, and artistic style expression.

- Date: 2025-04-15
- Lab: Kuaishou
- Type: Image
- Weights: Proprietary
- Modalities: text, image → image
- Available at launch: Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#kolors-2-0

Highlights:

- Launched on Kling AI alongside Kling 2.0 Master
- Improves prompt adherence, cinematic quality, and artistic styles

Source: [Kling AI Advances to the 2.0 Era, Empowering Everyone to Tell Great Stories with AI](https://ir.kuaishou.com/news-releases/news-release-details/kling-ai-advances-20-era-empowering-everyone-tell-great-stories)

### GLM-4-0414 and GLM-Z1

GLM-4-0414 is Z.ai's open-weights model series at 9B and 32B, released with the GLM-Z1 deep reasoning models and GLM-Z1-Rumination, which thinks longer on open-ended research tasks. GLM-4-32B-Base-0414 was pre-trained on 15T tokens.

- Date: 2025-04-14
- Lab: Z.ai
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Parameters: 9B, 32B
- License: MIT
- Available at launch: API, Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#glm-4-0414

Highlights:

- GLM-Z1-32B-0414 adds cold-start and extended reinforcement learning for math, code, and logic
- API launched GLM-4-Air-250414, GLM-4-Flash-250414, and GLM-Z1-Air, AirX, and Flash
- Weights under the MIT license

Source: [GLM-4 (GitHub)](https://github.com/zai-org/GLM-4)

### GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano

GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano are API models that improve on GPT-4o in coding and instruction following, with context windows up to 1 million tokens and a June 2024 knowledge cutoff.

- Date: 2025-04-14
- Lab: OpenAI
- Type: LLM
- Weights: Proprietary
- Modalities: text, image → text
- Context window: 1,047,576 tokens
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#gpt-4-1

Highlights:

- API model IDs gpt-4.1, gpt-4.1-mini, and gpt-4.1-nano
- 1,047,576-token context window for all three, per the API model pages
- GPT-4.1 nano is OpenAI's first nano model
- GPT-4.1 came to ChatGPT for all paid users on May 14, 2025

Source: [Introducing GPT-4.1 in the API | OpenAI](https://openai.com/index/gpt-4-1/)

### Nemotron-H

Family of hybrid base language models from NVIDIA that replace most attention layers with Mamba-2 layers to speed up inference, released under a research license.

- Date: 2025-04-14
- Lab: NVIDIA
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- License: NVIDIA Internal Scientific Research and Development Model License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#nemotron-h

Highlights:

- The 8B base model uses only four attention layers
- Described in the Nemotron-H technical report
- Base completion models intended for customization

Source: [nvidia/Nemotron-H-8B-Base-8K · Hugging Face](https://huggingface.co/nvidia/Nemotron-H-8B-Base-8K)

### Seed-Thinking-v1.5

Seed-Thinking-v1.5 is ByteDance Seed's reasoning model trained with large-scale reinforcement learning for math, coding, science, and creative writing; Volcano Engine serves it as Doubao-1.5-thinking-pro.

- Date: 2025-04-14
- Lab: ByteDance Seed
- Type: Reasoning
- Weights: Proprietary
- Modalities: text → text
- Parameters: MoE: 200B total, 20B active
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#seed-thinking-v1-5

Highlights:

- Mixture-of-experts with 200B total and 20B active parameters
- API access on Volcano Engine opened April 17, 2025
- Technical report published on GitHub

Source: [ByteDance's Latest Thinking Model, Seed-Thinking-v1.5 Technical Details Disclosed](https://seed.bytedance.com/en/blog/bytedance-s-latest-thinking-model-seed-thinking-v1-5-technical-details-disclosed)

### Kimi-VL

Kimi-VL is an efficient open-weights mixture-of-experts vision-language model for multimodal reasoning, long-context understanding, and agent tasks. It was released as Kimi-VL-A3B-Instruct and Kimi-VL-A3B-Thinking.

- Date: 2025-04-10
- Lab: Moonshot AI
- Type: Multimodal
- Weights: Open weights
- Modalities: text, image, video → text
- Context window: 128,000 tokens
- Parameters: MoE: 16B total, 3B active
- License: MIT
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#kimi-vl

Highlights:

- Activates about 2.8B parameters in its language decoder
- 128K extended context window
- Native-resolution MoonViT vision encoder for high-resolution inputs

Source: [moonshotai/Kimi-VL-A3B-Instruct (Hugging Face)](https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct)

### Amazon Nova Sonic

Amazon Nova Sonic is Amazon's first speech-to-speech foundation model, unifying speech understanding and generation in one model so developers can build real-time voice conversations without separate recognition and text-to-speech models.

- Date: 2025-04-08
- Lab: Amazon
- Type: Speech
- Weights: Proprietary
- Modalities: text, audio → text, audio
- Context window: 300,000 tokens
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#amazon-nova-sonic

Highlights:

- Launched with American and British English, adding Spanish in June 2025 and French, Italian and German in July 2025
- 300K token context window
- Launched in Amazon Bedrock in US East (N. Virginia)

Source: [Introducing Amazon Nova Sonic: Human-like voice conversations for generative AI applications](https://aws.amazon.com/blogs/aws/introducing-amazon-nova-sonic-human-like-voice-conversations-for-generative-ai-applications/)

### Amazon Nova Reel 1.1

Amazon Nova Reel 1.1 updates Amazon's video generation model to produce multi-shot videos of up to two minutes from 6-second shots, with style consistency across shots and improved single-shot quality and latency.

- Date: 2025-04-07
- Lab: Amazon
- Type: Video
- Weights: Proprietary
- Modalities: text, image → video
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2025/04#amazon-nova-reel-1-1

Highlights:

- Generates videos up to 2 minutes long
- Automated mode from one prompt or manual storyboard mode with up to 20 shots
- Available in Amazon Bedrock in US East (N. Virginia)

Source: [Amazon Nova Reel 1.1: Featuring up to 2-minutes multi-shot videos](https://aws.amazon.com/blogs/aws/amazon-nova-reel-1-1-featuring-up-to-2-minutes-multi-shot-videos/)

### Gen-4 Turbo

Faster and more efficient version of Gen-4, described by Runway as its fastest and most efficient video model at launch and made available to all users.

- Date: 2025-04-07
- Lab: Runway
- Type: Video
- Weights: Proprietary
- Modalities: text, image → video
- Available at launch: Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#runway-gen-4-turbo

Highlights:

- Available to all plans at launch
- Turbo variant of Gen-4

Source: [Product Updates & Changelog | Runway AI](https://runway.com/changelog)

### Llama-3.1-Nemotron-Ultra-253B-v1

Largest Llama Nemotron reasoning model, derived from Meta Llama-3.1-405B-Instruct with neural architecture search to cut memory use while keeping accuracy.

- Date: 2025-04-07
- Lab: NVIDIA
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Parameters: 253B
- License: NVIDIA Open Model License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2025/04#llama-3-1-nemotron-ultra-253b

Highlights:

- Post-trained for reasoning, chat preferences, RAG and tool calling
- Trained between November 2024 and April 2025
- Part of the Llama Nemotron collection

Source: [nvidia/Llama-3_1-Nemotron-Ultra-253B-v1 · Hugging Face](https://huggingface.co/nvidia/Llama-3_1-Nemotron-Ultra-253B-v1)

### Llama 4 Maverick

Llama 4 Maverick is a natively multimodal mixture-of-experts model with 17B active parameters and 128 experts, released with Llama 4 Scout as the first Llama 4 models.

- Date: 2025-04-05
- Lab: Meta
- Type: Multimodal
- Weights: Open weights
- Modalities: text, image → text
- Context window: 1,000,000 tokens
- Parameters: MoE: 400B total, 17B active, 128 experts
- License: Llama 4 Community License
- Available at launch: Open weights, Cloud platforms, Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#llama-4-maverick

Highlights:

- 400B total parameters; runs on a single NVIDIA H100 DGX host
- Model card lists a 1M-token context window and 12 supported languages
- Knowledge cutoff of August 2024
- Powers Meta AI in WhatsApp, Messenger, Instagram Direct and meta.ai

Source: [The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation](https://ai.meta.com/blog/llama-4-multimodal-intelligence/)

### Llama 4 Scout

Llama 4 Scout is a natively multimodal mixture-of-experts model with 17B active parameters and 16 experts, notable for a 10 million token context window.

- Date: 2025-04-05
- Lab: Meta
- Type: Multimodal
- Weights: Open weights
- Modalities: text, image → text
- Context window: 10,000,000 tokens
- Parameters: MoE: 109B total, 17B active, 16 experts
- License: Llama 4 Community License
- Available at launch: Open weights, Cloud platforms
- Permalink: https://theaiengineering.com/releases/2025/04#llama-4-scout

Highlights:

- 109B total parameters; fits on a single NVIDIA H100 GPU with Int4 quantization
- Accepts multilingual text and image input
- Llama 4 Behemoth was previewed but not released because it was still training

Source: [The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation](https://ai.meta.com/blog/llama-4-multimodal-intelligence/)

### Motion 2.0

Leonardo.Ai's own video generation model, built into its AI Creation interface to make videos from a text prompt or from images, with boosted frame rates for smoother playback.

- Date: 2025-04-04
- Lab: Leonardo.Ai
- Type: Video
- Weights: Proprietary
- Modalities: text, image → video
- Available at launch: Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#leonardo-motion-2-0

Highlights:

- Launched for paid users on web and mobile
- 720p output added on May 6, 2025
- Offered on Replicate with Phoenix 1.0 from June 10, 2025

Source: [Leonardo.Ai on X: Double Drop at Leonardo.Ai. Motion 2.0 is here](https://x.com/LeonardoAi/status/1908089989210018111)

### Midjourney V7

Seventh-generation image model with better text and image prompt handling, higher image quality and coherence, personalization on by default, and a new Draft Mode that renders at ten times the speed.

- Date: 2025-04-03
- Lab: Midjourney
- Type: Image
- Weights: Proprietary
- Modalities: text, image → image
- Available at launch: Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#midjourney-v7

Highlights:

- Launched as an alpha with Turbo and Relax modes
- Became the default model on June 17, 2025
- Introduced Omni Reference and Draft Mode

Source: [Version - Midjourney](https://docs.midjourney.com/hc/en-us/articles/32199405667853-Version)

### MiniMax Speech-02

Speech-02 is MiniMax's text-to-audio and voice cloning model series, offered as Speech-02-HD for high-fidelity voiceovers and audiobooks and Speech-02-Turbo for low-latency interactive use.

- Date: 2025-04-02
- Lab: MiniMax
- Type: Speech
- Weights: Proprietary
- Modalities: text → audio
- Available at launch: API, Consumer app
- Permalink: https://theaiengineering.com/releases/2025/04#minimax-speech-02

Highlights:

- Voice cloning from 10 seconds of recorded audio
- Library of more than 300 prebuilt voices
- Built on an autoregressive Transformer with a learnable speaker encoder, per MiniMax's May 2025 technical post

Source: [Speech-02-series : The Next Leap in Text-to-Audio and AI Voice Technology](https://www.minimax.io/news/speech-02-series)

## Cite and reuse this data

These release records are free to reuse under CC BY 4.0, including commercially. Credit The AI Engineering and link to this page's URL. Dates are the labs' official announcement dates, and every record links to its source. List fields in the CSV are separated by " | ".

The AI Engineering Editors. "AI models released in April 2025." The AI Engineering, updated Apr 30, 2025. https://theaiengineering.com/releases/2025/04

HTML link:

```html
<a href="https://theaiengineering.com/releases/2025/04">AI models released in April 2025</a> (The AI Engineering)
```

- Title: AI models released in April 2025
- Author: The AI Engineering Editors
- Publisher: The AI Engineering
- Updated: 2025-04-30
- URL: https://theaiengineering.com/releases/2025/04
- Markdown: https://theaiengineering.com/releases/2025/04.md
- License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
- Full dataset as CSV: https://theaiengineering.com/releases.csv
- Full dataset as JSON: https://theaiengineering.com/releases.json
