According to DeepLearningAI, AREX verifies requirements, scores 82.5 BrowseComp and 82.0 F1 WideSearch en, and its harness alone adds up to 10 points.

According to @TheRundownAI, Anthropic’s Claude Haiku 5.5 is faster, cheaper, and tops Luna on benchmarks, starting at $0.10 per million input tokens.

According to @timnitGebru, OpenAI researchers said GPT3 achieved AGI in 2020, raising scrutiny of benchmarks and model evaluation standards.

According to @emollick, a new visual should replace METR’s Long Task Horizon chart, prompting fresh analysis of AI capability benchmarks.

According to DeepLearningAI, Meta pairs action and memory agents to curb context rot, boosting Claude Sonnet 4.5 accuracy from 37.6% to 45.9%.

According to @emollick, public AI benchmarks are saturated or error ridden, severely underestimating current model abilities, as the cited paper shows.

According to DeepLearning.AI, Claude Fable 5.1 leads Intelligence Index v4.2 with 57, ties OpenAI’s new model on benchmarks, and cuts agent costs via cache.

According to KyeGomezB, DeepSeek’s causal encoder‑decoder MoE cuts active params to 8B in, 16B out, promising lower cost and top benchmarks.

According to StanfordAILab, Terminal-Bench-Science v0.1 tests 70 research tasks and Claude Opus 5 solves about 30%, signaling big gaps in agent tooling.

According to Ethan Mollick, Qwen 27B is a game changer, but critics on X urge benchmarking, noting weaker agentic task performance.

According to gdb, OpenAI’s GPT 5.x shows major gains on internal personal finance benchmarks, pointing to stronger planning and advisory performance.

According to emollick, a Codex-prompted project produced BBBBB, an executable benchmark testing AI-authored conformance suites for metrics.

According to TheRundownAI, Anthropic’s Claude Opus 5 leaps past 4.8 and outperforms Fable on several benchmarks at half the price.

According to @CNBC, Moonshot AI says Kimi K3 beats top OpenAI and Anthropic models on select benchmarks while still trailing overall leaders.

According to The Rundown AI, Grok 4.5 in SpaceXAI and Cursor delivers flash-level speed, strong benchmarks, and affordable pricing, boosting dev workflows.

According to StanfordAILab, Stanford AI Lab lists ICML 2026 papers on coding agents, LLM reasoning, safety, interpretability, and science.

According to CNBC... Zhipu’s GLM models gain on OpenAI and Anthropic as open source and export limits reshape AI competition, per benchmarks and funding data.

According to OpenAI... LifeSciBench debuts with 750 expert tasks across 7 workflows to assess AI for real-world life science research.

According to OpenAI on X, leaders discuss better evals to forecast model progress as saturated benchmarks get gamed, outlining next judgment areas.

According to God of Prompt, Anthropic is crushing a new benchmark, signaling Claude3.5 gains for reasoning and eval leadership.