Artificial Analysis launched version 5 of its Intelligence Index, a composite benchmark measuring AI model capabilities, adding two new tests (FrontierCode 2.0 and Terminal-Bench-Science) and upgrading three others, including a refined version of Humanity's Last Exam. The firm also introduced a companion Progress Index to track model capability over time, and reports that 60% of the Index now relies on private or held-out test data. In current results, Claude Opus 5.5 leads with a score of 58, narrowly ahead of GPT-6 Astra at 57.
artificialanalysis.ai
· 2026-10-28
Moonshot AI released Kimi K3.1 Pro, an open-weight 3T-class model built on Kimi K3 with expanded post-training compute, offering a 1-million-token context window and gains in coding, 3D creation, financial research and agentic tasks. A faster variant, K3.1 Pro Highspeed, runs 2-3 times quicker and both are live now on Kimi.ai, Kimi Work, Kimi Code and the Kimi API, with full model weights due by November 4.
kimi.ai
· 2026-10-23
DeepSeek has launched V4.1-Flash-1017, a minor version update to its V4.1-Flash model that the company says improves reasoning, coding, agentic task handling, and visual benchmark scores compared to the prior version. The update also shortens task execution time and output length. DeepSeek is also releasing deepseek-recipe, a Rust-based toolkit with Python bindings for encoding prompts and parsing model outputs.
huggingface.co
· 2026-10-17
Weave OS released Weave Router 2.0, an open-source model router that dynamically switches between LLMs for coding agents like Claude Code or Codex. On Terminal Bench 4.0 and SWE Atlas benchmarks, the router matched GPT-6 Astra's pass rates while running at roughly 52-54% of its cost and 2.2-2.5x faster. The company attributes the gains to a new routing architecture, a larger training dataset, and improved cache-eviction calculations.
news.ycombinator.com
· 2026-09-30
Anthropic released Claude Sonnet 5.5, the second model in its Claude 5.5 lineup after Opus 5.5. The company says it runs over 30% faster and costs up to 30% less per task than Sonnet 5, while keeping the same per-token pricing, and jumps from 10.3% to 70.6% on the Terminal-Bench 4.0 agentic coding evaluation. A third model, Claude Haiku 5.5, is planned for release in the coming weeks.
anthropic.com
· 2026-09-28
Anthropic released Opus 5.5, an update to Claude aimed at enterprise tasks like coding, financial analysis and business work, claiming better agentic coding benchmarks than GPT-6 Astra and reduced token pricing compared to Opus 5. OpenAI simultaneously released GPT-6 Sol and GPT-6 Luna, cheaper successors to GPT-5.6 that the company says cut factual errors roughly in half and match rival Fable 5.1's coding performance at lower cost.
engadget.com
· 2026-09-22
RTK, a popular tool with over 79,000 GitHub stars that compresses terminal output before AI coding agents read it, has been marketed as a way to cut AI coding costs, with one viral post claiming up to 60% token savings. But independent benchmark testing using Terminal-Bench 2.1 across 1,740 task attempts found mixed results: costs dropped 5% for one model setup but rose 5% for another, contradicting the widely shared savings figures.
quesma.com
· 2026-09-11
Cognition released SWE-2, a coding model built by post-training on a 2.8-trillion-parameter Kimi K3 base with 104B active parameters, marking its first RL push into the multi-trillion-parameter range. The model posts a class-leading 92.8 on Terminal-Bench 2.1 and competitive FrontierCode scores at lower cost than rivals like Claude Fable 5.1 and GPT-6 Astra, but falls well behind both on the harder Terminal-Bench 4.0 benchmark (27.3 vs 55.8 and 57.9). SWE-2 is closed-weight and available now through Devin Desktop and CLI, with no local deployment option.
tokenstead.ai
· 2026-09-10
A benchmarking study tested several GGUF quantizations of the Qwen3.8 27B model against benchmarks including GPQA Diamond, IFBench, and Terminal-Bench 2.1. The 17GB Q4_K_M version performed on par with the full 55GB BF16 model on coding tasks and fits on a consumer 24GB GPU, while the smallest 6.2GB 1-bit version dropped to near-random accuracy on reasoning tasks, with longer reasoning chains making results worse.
quesma.com
· 2026-09-08
Stanford researchers, working with the Terminal-Bench team and scientists across disciplines, launched Terminal-Bench-Science, a benchmark testing AI agents on real scientific research workflows. The initial version includes 70 tasks spanning life, physical, Earth, mathematical, and engineering sciences, with Claude Opus 5 running Claude Code topping the leaderboard at a 30% resolution rate.
terminal-bench-science.ai
· 2026-08-28
GoKawiil