Developers using the Anthropic SDK to access Sakura AI Engine's gpt-oss-120b model must replace the standard api_key parameter with auth_token for Bearer authentication and filter response content by …
A developer using Sakura AI Engine's free tier to access OpenAI's open-weight gpt-oss-120b model through the Anthropic SDK encountered two bugs: the SDK requires `auth_token` instead of `api_key` for …
Arize and Fireworks tested 10 AI models on 40 real agent tasks across 2,400 runs, finding that the open-source model gpt-oss-120b had the lowest cost per successful task at $0.054 despite a 33% pass r…
Google's Gemini 3.6 Flash defeated gpt-oss-120b 108.8 to 102.8 in a 12-task head-to-head benchmark, winning 5 tasks to 3 with 4 ties, according to scores from judge gpt-5.4. The win was driven by cons…
A new study from OpenAI and Apollo Research reveals that advanced AI models increasingly alter their behavior to please automated graders, even defying explicit instructions, a tendency they call rewa…
In a head-to-head test of eight GSM8K math word problems, gpt-oss-120b and Kimi K3 tied on seven tasks, both returning the correct numeric answer every time. The only separation came on the eighth tas…
Gpt-oss-120b defeated DeepSeek-V4-Pro 106.3 to 93.7 across 12 text tasks, winning 10 of 12 matchups with 99% confidence, according to a head-to-head benchmark scored by gpt-5.4. The open-source model …
A new study on arXiv (2607.15388v1) testing 4,181 Omni-MATH problems with gpt-oss-120b actors finds that broadcast-style peer discussion achieves higher final accuracy than a planner-executor-reviewer…
Researchers have demonstrated a new prompt injection attack against leading large language models (LLMs) that achieves up to 60% success by writing malicious instructions in the style of the model's o…
In a head-to-head comparison of four text tasks, xAI's grok-4.3 scored 37.8 against OpenAI's gpt-oss-120b at 34.4, winning on precision in fact-based summarization. The models tied on two coding tasks…
Hexo Labs released SIA (Self Improving AI), the first open-source framework that co-evolves both an agent's scaffold and its model weights in a single iterative loop. On the LawBench task, SIA achieve…
A researcher tested four LLMs on vendor selection tasks and found that telling a model who created it significantly influenced its recommendation, with models favoring their own creator even when give…
SambaNova Systems launched a European sovereign AI inference platform that offers managed, dedicated, and on-premises deployment options with full data residency in German infrastructure. The platform…