#model-release

26 articles

Claude Fable 5 Just Shipped: 80.3% on SWE-Bench Pro, 2× Opus 4.8 Pricing, Free Through June 22

Anthropic released the first publicly available Mythos-class model today — Claude Fable 5, with safeguards that fall back to Opus 4.8 on high-risk prompts. 80.3% SWE-Bench Pro (11 points over Opus 4.8). $10/$50 per 1M tokens. Free on Pro/Max plans until June 22, then credit-metered.

2026-06-0910 min read

Claude Sonnet 4.8: What the Leak Actually Says, and Why the Pattern Doesn't Fit

A 'Sonnet 4.8' string appeared in a Claude Code source-map leak on March 31. Anthropic has shipped nothing. Polymarket is at 3% for today. And a version jump from 4.6 → 4.8 (skipping 4.7) would be unprecedented for Anthropic. Here's the honest read.

2026-05-247 min read

Seedance 2.1 and Seedance 2.0 Mini Are Coming: Quality Bump, Lower Price Tier

ByteDance is preparing two new Seedance variants: Seedance 2.1 with a reported ~20% generation-quality improvement, and a new Mini tier said to come in well below the current Seedance pricing while outperforming Seedance 2.0 Fast.

2026-05-197 min read

GPT-5.6 Just Showed Up in OpenAI's Codex Logs — Here's What That Actually Means

A single routing entry in OpenAI's Codex rollout logs pointed at GPT-5.6. Polymarket is at 89% on a June 30 release. Here's what's confirmed, what's noise, and why the goblin incident is the reason testing is moving this fast.

2026-05-148 min read

HiDream-O1-Image-Dev: The 8B Pixel-Native Model That Beat 56B FLUX.2

HiDream-O1-Image-Dev is an 8B distilled image model that drops the VAE and the external text encoder, generates 2K natively, and outscores models 7x its size on GenEval, DPG, and HPSv3.

2026-05-108 min read

FLUX.2 Official Launch — Next-Generation Visual AI Arrives on WaveSpeedAI

We are thrilled to announce the official launch of FLUX.2, the newest flagship model from Black Forest Labs (BFL), now available on WaveSpeedAI. It represents a major leap forward in visual AI, delivering unprecedented levels of prompt understanding, visual fidelity, and creative flexibility.

2025-06-112 min read

Flux.2 Arrives on WaveSpeedAI: Honest Comparison with NanoBanana Pro

Is Flux.2 still competitive in 2025? With the overwhelming attention Nano Banana Pro is receiving, this article gives you an honest, prompt-based breakdown.

2025-06-074 min read

LongCat Avatar Is Live on WaveSpeedAI: Ultra-Realistic Lip-Synced Avatar Videos Up to 2 Minutes

LongCat Avatar transforms a single photo and an audio track into super-realistic, lip-synchronized talking or singing avatar videos, with natural dynamics and consistent identity—for up to 2 minutes per generation.

2025-05-064 min read

Seedream4.5 Launches on WaveSpeedAI: A Major Leap in Visual Generation Performance

WaveSpeedAI has officially integrated Seedream4.5 by ByteDance, bringing one of the most advanced multimodal image-generation models to its platform. Seedream4.5 delivers sharper detail, higher stability, and more accurate interpretation of complex prompts.

2025-04-185 min read

FLUX.1 Kontext Dev Now Live on WaveSpeedAI: Open-Weight Editing with Precision, Speed, and Contextual Intelligence

2025-03-173 min read

FLUX.1 Kontext Now Live on WaveSpeedAl: Professional AI Image Editing

We're excited to bring you FLUX.1 Kontext, a cutting-edge image editing model from Black Forest Labs, now available on WaveSpeedAI.

2025-03-132 min read

FLUX-dev Arrives on WaveSpeedAI: 12B Parameter AI Model Redefining Text-to-Image Generation

FLUX-dev, the latest model launched in WaveSpeedAI, has new creative possibilities in text-to-image generation with 12B parameters and open-source flexibility. This blog will help you learn more about the model and its potential.

2025-03-114 min read

Four New MiniMax Hailuo 02 Video Models are Now Live on WaveSpeedAI

WaveSpeed AI is excited to announce the addition of four new models from MiniMax’s acclaimed Hailuo 02 series: t2v-pro, i2v-pro, i2v-standard, and t2v-standard.

2025-03-103 min read

Ghibli Now Live on WaveSpeedAI

Discover the groundbreaking Ghibli model on WaveSpeedAI, enabling high-quality video generation with unprecedented ease and efficiency. Explore its features, use cases, and why WaveSpeedAI is the ideal platform for your creative needs.

2025-03-075 min read

HiDream-E1-Full Now Available on WaveSpeedAI

We’re excited to announce the launch of HiDream-E1-Full, a powerful image editing model from HiDream AI, now available on WaveSpeedAI.

2025-03-052 min read

HiDream-I1 Now Live on WaveSpeedAI: The Next-Gen Open-Source Image Generation Model

WaveSpeedAI is excited to announce support for HiDream-I1, a groundbreaking open-source image generative foundation model with 17 billion parameters. This state-of-the-art model delivers exceptional image quality and prompt adherence, making it ideal for creators, researchers, and enterprises seeking high-performance generative AI solutions.

2025-03-043 min read

Hunyuan3D 2.0 Now Live on WaveSpeedAI: Revolutionizing High-Resolution Textured 3D Asset Generation

WaveSpeedAI is excited to announce support for Hunyuan3D 2.0, a groundbreaking open-source 3D generation model with 17 billion parameters. This state-of-the-art model delivers exceptional 3D asset generation, making it ideal for creators, researchers, and enterprises seeking high-performance generative AI solutions.

2025-03-014 min read

HunyuanCustom Now Live on WaveSpeedAI: Rivals Top Closed-Source Solutions!

We’re excited to announce HunyuanCustom, An open-source, multimodal-driven architecture for customized video generation, now available on WaveSpeedAI.

2025-02-282 min read

InstantCharacter Model Launches on WaveSpeed: Ushering in a New Era of Personalized Character Generation

We’re excited to announce that Tencent AI Lab’s latest innovation — InstantCharacter, a state-of-the-art personalized character generation model — is now officially live on the WaveSpeed platform. Built upon a scalable Diffusion Transformer framework, InstantCharacter offers high fidelity, strong generalization, and fine-grained text controllability, setting a new benchmark in character generation technology.

2025-02-252 min read

Kling V2.1 Now Live on WaveSpeedAI: Cinematic Video Generation

2025-02-102 min read

Magi-1 Now Live on WaveSpeedAI: A New Benchmark in Open-Source Video Generation

Magi-1, the groundbreaking open-source video generation model by Sand AI, is now available on WaveSpeedAI for real-time inference and API deployment.

2025-02-093 min read

MiniMax Video-01 Now Live on WaveSpeedAI

Discover the groundbreaking MiniMax Video 01 model on WaveSpeedAI, enabling high-quality video generation with unprecedented ease and efficiency. Explore its features, use cases, and why WaveSpeedAI is the ideal platform for your creative needs.

2025-02-064 min read

MMAudio V2 Now Live on WaveSpeedAI: Multimodal Audio Generation for Video and Text Inputs

MMAudio V2, a cutting-edge model designed to generate high-quality, synchronized audio from both video and text inputs. This advancement opens new possibilities in multimedia content creation, enhancing the realism and immersion of generated media

2025-02-042 min read

Real-ESRGAN Now Live on WaveSpeedAI: Revolutionizing Image Enhancement

Real-ESRGAN, a powerful tool for enhancing image quality, now available on WaveSpeedAI. Developed by Tencent's ARC Lab, this model excels at transforming low-resolution images into high-resolution masterpieces, even when the exact nature of the image degradation is unknown.

2025-02-012 min read

Video Upscaler Now Live on WaveSpeedAl: The Best Model for Improving Video Resolution

We are excited to introduce the video-upscaler, a powerful tool designed to enhance the resolution and quality of videos, now live on WaveSpeedAI!

2025-01-162 min read

Vidu 2.0 Now Live on WaveSpeedAI: A New Generation of Controllable Video Generation Models

We’re excited to introduce the full suite of Vidu 2.0 models—now available on WaveSpeedAI. Designed by Shengshu Technology, these models bring a new level of controllability, visual fidelity, and temporal consistency to open-source video generation.

2025-01-153 min read