AiBattle
@AiBattle_
Artificial Intelligence
Kimi K2 🆚 Qwen-3-235B-A22B-2507 The new updated Qwen 3 model beats Kimi K2 on most benchmarks. The jump on the ARC-AGI score is especially impressive An updated reasoning model is also on the way according to Qwen researchers

Google's advanced DeepThink model also achieved gold at the IMO using natural language and within the given time limit A version of the DeepThink model that achieved gold will be available to Google AI Ultra subscribers, possibly much earlier than the model used by OpenAI
Very excited to share that an advanced version of Gemini Deep Think is the first to have achieved gold-medal level in the International Mathematical Olympiad! 🏆, solving five out of six problems perfectly, as verified by the IMO organizers! It’s been a wild run to lead this…
We could see GPT-5 released before September OpenAI pulled the "o3-Alpha" model from public testing just 12 hours after it went live, maybe an indication that a full launch is close. In the past, when OpenAI tested secret models like "Optimus Alpha" and "Quasar Alpha," the…
Heard GPT-5 is imminent, from a little bird. - It’s not one model, but multiple models. It has a router that switches between reasoning, non-reasoning, and tool-using models. - That’s why Sam said they’d “fix model naming”: prompts will just auto-route to the right model. -…
Google Gemini Checkpoint / Model Summary – July update Changes: - The model "Wolfstride" was added on July 4.

Several new models have entered LmArena for testing: - Clownfish: Claims to be Deepseek R1 - Nettle: Claims to be Deepseek R1 - Octopus: Claims to be Deepseek R1 - Cresylux: Claims to be developed by Meituan Will we need to do some testing to find out the what these models…

For those dissatisfied with the speed of Kimi K2 on the website, @GroqInc currently hosts K2 at over 230 T/s, 20 times faster than most providers
Kimi K2 non-reasoning is already great at coding and creative writing, can't wait to see the performance of K2 with reasoning
Kimi K2 non-reasoning is already great at coding and creative writing, can't wait to see the performance of K2 with reasoning

Kimi K2 - 3D model of an AK-47 K2's habit of adding particle effects reminds me of the Claude models.