- The Summary AI
- Posts
- 🦙 Meta Releases Llama 4
🦙 Meta Releases Llama 4
PLUS: Free ChatGPT Plus for College Students
The Summary AI
April 07, 2025
Welcome back!
Meta has released its Llama 4 open-source models, outperforming GPT-4.5 in key benchmarks and offering a 10-million-token context window. It’s a bold push to make cutting-edge AI more accessible, and it's already shaking up the leaderboards. Let’s unpack…
Today’s Summary:
🔥 Meta drops Llama 4 models
🎬 Runway debuts Gen-4 video AI
🎓 ChatGPT Plus free for US college students
🛒 Amazon’s AI shops for you
🖼️ Midjourney launches version 7
🧐 AI models hide their true thoughts
🛠️ 2 new tools
TOP STORY
Meta releases Llama 4
The Summary: Meta has launched two powerful open-weight AI models, Llama 4 Scout and Maverick, pushing multimodal open-source AI into frontier territory. Scout (109B) fits on a H100 GPU and supports a record-breaking 10M-token context window. Maverick ranks #2 on LM Arena, beating GPT-4o, GPT-4.5 and Grok, with only Google’s Gemini 2.5 Pro ahead. Both are distilled from the still-training 2T parameters “Behemoth” model, Meta’s strongest move yet to stay in the frontier race.
Key details:
#1 open model on LM Arena, outperforming GPT-4.5 and Grok 3 on hard prompts, coding, math, and creative writing
Multimodal, with native support for image input
Scout’s 10M token context window is the longest of any major model and allows ingestion of entire codebases
Trained on 30T tokens with 10× more languages than Llama 3
Costs from $0.15 per 1M tokens, x1000 cheaper than GPT-4.5
Not licensed for EU developers due to regulatory uncertainties, but end users are allowed
Requires H100 GPUs, but its MoE design makes it viable on high-memory Macs, since only a subset of parameters runs per token
Why it matters: Meta’s release is the strongest open-source challenge yet to proprietary giants like OpenAI. With Scout’s 10M context length and Maverick’s efficiency, open models are now competing at frontier level, not just catching up.
RUNWAY AI
Runway unveils Gen-4, its most advanced video AI yet
The Summary: Runway released Gen-4, its newest AI video model built to fix a key problem in AI filmmaking: visual consistency across scenes. With character, object, and environment coherence, Gen-4 allows actual story-driven content without weird glitches or shifting faces. Available to paid users.
Key details:
Gen-4 maintains consistent characters, objects, and environments using a single reference image
Users can generate 5–10 second clips at 720p with multi-angle, style-consistent scenes
Gen-4 was used to create short films like New York is a Zoo and The Retrieval, made in under a week
Why it matters: Gen-4 pushes AI video past novelty into actual production. Until now, AI-generated clips looked cool but fell apart across scenes, with shifting faces, vanishing props, and weird physics. Thanks to Gen-4 visual continuity, it becomes a valuable professional tool that understands scenes and characters.
FROM OUR PARTNERS
Writer launches workplace-ready agentic AI
You’ve heard the hype. It’s time for results.
After two years of siloed experiments, proofs of concept that fail to scale, and disappointing ROI, most enterprises are stuck. AI isn't transforming their organizations — it’s adding complexity, friction, and frustration.
But Writer customers are seeing positive impact across their companies. Our end-to-end approach is delivering adoption and ROI at scale. Now, we’re applying that same platform and technology to build agentic AI that actually works for every enterprise.
This isn’t just another hype train that overpromises and underdelivers. It’s the AI you’ve been waiting for — and it’s going to change the way enterprises operate. Be among the first to see end-to-end agentic AI in action. Join us for a live product release on April 10 at 2pm ET (11am PT).
Can't make it live? No worries — register anyway and we'll send you the recording!
OPENAI
ChatGPT Plus goes free for students through May
The Summary: OpenAI is giving college students in the U.S. and Canada free access to ChatGPT Plus through May 2025. The offer lands during finals season and opens up GPT-4o, image generation, voice chat, and Deep Research tools to millions of students. The move comes just a day after Anthropic launched its own student-focused AI suite.
Key details:
ChatGPT Plus is now free for all college students in the US and Canada until May 31, 2025
Students already subscribed to Plus also get two months free
The offer is verified through SheerID using trusted sources
Claude for Education, launched one day earlier, includes a “Learning Mode” with Socratic guidance
1 in 3 U.S. adults aged 18-24 use ChatGPT; a quarter of their queries are school-related
Why it matters: OpenAI’s free offer lands as finals season hits, giving students free access when they need it most. If ChatGPT proves genuinely helpful now, it’s more likely to stick and shape how the next generation learns, creates, and works.
QUICK NEWS
Quick news
Amazon “Buy for me” agent can shop from other retailers
Midjourney releases v7 image model
TOOLS
🥇 New tools
That’s all for today!
If you liked the newsletter, share it with your friends and colleagues by sending them this link: https://thesummary.ai/