DeepSeek released two powerful AI models, DeepSeek-V3.2 and DeepSeek-V3.2-Speciale, matching the capabilities of GPT-5. These models can be used for free under an open-source MIT license.
Why it matters
DeepSeek's release of two powerful AI models that can match the capabilities of GPT-5 marks a significant development in the field of AI and challenges the dominance of American technology giants.
Community talk
[Release] We built Step-Audio-R1: The first open-source Audio LLM that truly Reasons (CoT) and Scales – Beats Gemini 2.5 Pro on Audio Benchmarks.
🤩 Deep Research Tulu (DR Tulu) now beats Gemini 3 Pro on key benchmarks
OpenAI's Mark Chen reaction to Gemini 3
Gemini app downloads are catching up to ChatGPT
New Model Step-Audio-R1 open source audio model to actually use CoT reasoning, close to Gemini 3
Gemini 3 Feature - Paranoid Delusions
Gemini has added AI photo detection, upload any image and just ask.
Built a tool that pits Claude Opus against GPT and Gemini to stress-test ideas
Garlic: OpenAI’s New Small Model That Beats Gemini 3 & Opus 4.5 in Coding Benchmarks
LMArena update: Claude Opus 4.5 wins the “Triple Crown” and passes Gemini 3 Pro
for agentic Opus 4.5 >>> gemini 3 : for multimodel super reasoning gemini 3 >>>opus 4.5
Comparing GPT-5.1 vs Gemini 3.0 vs Opus 4.5 across 3 coding tasks. Here's an overview
"Gemini 3 Pro is the best model yet"
Raw Chain-of-Thought from Gemini 3 Pro. It hallucinates, corrects itself, and eventually crashes.
I just tried Google Gemini 3 voice mode and it is far better than ChatGPT. OpenAI, you have to raise your game.
gpt-5.1-search is superior to Gemini-3-pro-grounding
“Benchmarking” GPT-5.1, Gemini 3 Pro, and Opus 4.5 on designing emotional states for 12B Local Models (Unity Integration)
Experiences with ChatGPT5.1 vs. Gemini 3 pro
Amp Code switches to Opus 4.5 a week after the "historic" switch to Gemini 3.0
Ultimate Custom Instructions - Gemini and GPT
The METR evals for Gemini 3.0 and Opus 4.5 are taking incredibly long--GPT 5.1 codex max was benchmarked almost instantly as well as others. Why is that?
Gemini 3 Pro's updated IQ test results have declined.
FLUX.2 vs. Qwen Image Edit 2509 vs. Gemini 3 Pro Image Preview
Experience Report from Gemini 2.5 to Opus 4.5