Models

Google lists Nano Banana 2.1 as generally available in the Gemini API from October 6, with 1:8 and 8:1 aspect ratios and image output at $30 per million tokens.

OpenAI's Decisions API, in beta since October 6, returns probabilities, choices and scores from gpt-6-luna; OpenAI lists input at $0.10 per 1M tokens.

Anthropic's Claude Haiku 5.5 lists at $0.10 per million input tokens for prompts up to 100,000 tokens, versus $1 for Haiku 4.5, and adds a 1M-token context window and adaptive thinking.

Mistral AI launched Mistral Large 4, a 1-trillion-parameter multimodal model with 49 billion active parameters, in public preview, with open weights promised after safety testing.

Aleph Alpha published Kolibri on Hugging Face on October 3 under Apache 2.0; its own tech report puts the model on the quality-versus-serving-cost Pareto frontier.

Reka released a research preview of Rho-1, a 19-billion-parameter model that handles text, image, video and robot-action generation in one network, trained on 320 H100 GPUs.

Researchers from CMU, NYU, Stanford and MIT report in Nature that Ataraxos beat Pim Niemeijer with 15 wins, 1 loss and 4 draws, using far less compute than DeepMind's DeepNash.

Google announced Gemini 4 Argon, which it says leads on 12 of 18 disclosed benchmarks. Access begins with trusted cyber defenders, with introductory pricing of $2 and $10 per million tokens.

OpenAI unveiled GPT-6.1 Sol at DevDay, saying it nearly matches GPT-6 Astra on agentic coding at one-fifth the token prices; a GPT-6.1 Astra release was scrapped.

Sonnet 5.5 arrives in Claude Code at unchanged $2/$10 token pricing; Anthropic reports 70.6% on Terminal-Bench 4.0, up from 10.3% for Sonnet 5.

Xiaomi's MiMo-V2.6-Pro scored 46 on Artificial Analysis' Intelligence Index, the top open-weight result, with MIT-licensed weights and low API pricing.

Two independent researchers used OpenAI's Astra and Anthropic's Claude Opus 5 to decrypt German Army Enigma messages that had resisted cryptanalysis for decades.