Terminal-Bench-Science
PulseAugur coverage of Terminal-Bench-Science — every cluster mentioning Terminal-Bench-Science across labs, papers, and developer communities, ranked by signal.
- 2026-08-28 research_milestone A new benchmark, Terminal-Bench-Science, was released to evaluate AI agents on scientific research workflows. source
1 day(s) with sentiment data
-
Anthropic's Claude Opus 5.5 excels in video generation and benchmarks · 2 sources tracked
Anthropic has released Claude Opus 5.5, which has quickly garnered positive attention, particularly for its capabilities in generating explainer videos. The model demonstrates strong performance on various benchmarks, i…
-
OpenAI launches GPT-6 Astra, Anthropic releases Claude Fable 5.1 with new controls · 1 source tracked
OpenAI has released GPT-6 Astra, its first full version-number jump in over a year, which the company claims is state-of-the-art in computer use, browser use, and software engineering, with its president suggesting it m…
-
Anthropic cuts Fable 5.1 cache costs, making agentic AI cheaper · 2 sources tracked
Anthropic has released its latest models, Fable 5.1 and Mythos 5.1, with significant price reductions on cached input, a key factor for agentic AI applications. While headline prices remain the same, the cost for cache …
-
Anthropic launches Claude Fable/Mythos 5.1 with price cuts and enhanced autonomy
Anthropic has released its new flagship models, Claude Fable 5.1 and Claude Mythos 5.1, positioning them as the most advanced for coding and knowledge work. While maintaining the same per-token pricing as their predeces…
-
Anthropic launches Claude Fable 5.1 with improved coding, lower costs · 10 sources tracked
Anthropic has released its latest AI models, Claude Fable 5.1 and Mythos 5.1, which offer improved performance in coding and scientific research. Fable 5.1, the generally available version, shows significant gains on be…
-
New benchmark tests AI agents on 70 scientific research workflows
A new benchmark, Terminal-Bench-Science, has been released to evaluate AI agents on complex scientific research tasks. Developed by researchers at Stanford University and a global team of domain experts, it features 70 …