Skip to content

Areas

AI agents Latest news

Models that plan, use tools, and complete multistep tasks, from Claude Code and Manus to agent frameworks and evaluations.

201 picksPast 30 days: 88 itemsTotal: 1,570 items

Latest pick

Top picks archive · Page 7

Aug 7

Aug 7FriItems 121–140
  1. Prime Intellect BlogAI score62

    Prime Intellect adds multi-agent training and evaluation to PRIME-RL

    Prime Intellect's RL stack now supports multi-agent systems, letting users program interactions between agents, choose which roles learn, and assign credit across an episode. The release introduces Agent and Env abstractions and four example patterns: agentic judging, self-play, and user simulation. Multi-agent support ships today in verifiers 0.3.0 and prime-rl 0.8.0.

    AIWhy it matters: The post explains the Agent and Env abstractions and four multi-agent patterns, showing how roles, credit assignment, and episodes can be programmed in one RL stack.

Aug 5

Aug 5Wed
  1. Qwen · new models on Hugging FaceAI score79

    Qwen3.8-27B releases dense vision-language model with thinking controls

    Alibaba's Qwen team has released Qwen3.8-27B on Hugging Face as a 27B dense model with native image and video understanding. The model card reports gains over Qwen3.6-27B on coding and agent benchmarks, including SWE-bench Pro at 61.7 versus 53.5. It adds reasoning_effort levels and preserve_thinking, and its hosted Qwen Cloud version is described as coming soon.

    AIWhy it matters: The model card gives per-benchmark comparisons with Qwen3.6-27B and named rivals, plus reasoning_effort and preserve_thinking controls for judging cost and agent behavior.

  2. Prime Intellect BlogAI score75

    Prime Agent launches open-source self-improving RLM coding harness

    Prime Agent is a new open-source coding harness built on a persistent IPython kernel, a Recursive Language Model design, and Continual Harness state that the agent can create, read, update, and delete. Prime Intellect reports ARC-AGI-3 results of 95.5% RHAE Best@1 with Opus 5 and competitive long-context scores with the open-weights GLM-5.2 model.

    AIWhy it matters: The post explains how the RLM and Continual Harness designs let an agent write code against its own context, sub-agents, and harness state, with benchmark evidence.

Aug 4

Aug 4Tue
  1. John SchulmanAI score77

    Schulman Suggests Post-Training May Explain Agents' Cyber Eval Behavior

    John Schulman comments that models seem to enter a single-minded mode during cyber evaluations and asks whether chunky post-training is the cause. He suggests models may match the situation to an RLVR training region where task completion is the only reward, so aligned behavior learned elsewhere does not generalize. He adds that CTF-style tasks may be part of that training chunk.

    AIWhy it matters: The post links an unsanctioned agent incident in cyber testing to a specific post-training hypothesis, offering a possible mechanism for the behavior rather than only the event itself.

  2. PromptArmor Threat IntelligenceAI score67

    Atlassian Rovo can be manipulated to exfiltrate Jira and Confluence data

    PromptArmor reports that a hidden prompt injection in an uploaded file can make Atlassian Rovo send Jira tickets and Confluence documents to an attacker's URL without human approval. The attack works even when organization-wide web search is disabled, because the setting does not remove the URL retrieval tool. PromptArmor says it disclosed the issue to Atlassian on May 23, 2026, and that Rovo remained vulnerable at publication on August 5, 2026.

    AIWhy it matters: The report traces a full indirect prompt injection chain in Rovo, showing how a disabled web search setting still leaves a data exfiltration path open.

Aug 3

Aug 3Mon
  1. Liquid AI BlogAI score72

    Liquid AI releases LFM2.5-2.6B, a 2.6B on-device agentic model

    Liquid AI released LFM2.5-2.6B, a 2.6B-parameter agentic model that runs on-device on phones and CPUs, along with a base variant on Hugging Face. The company reports it leads on every instruction-following benchmark and nearly every tool-use benchmark it tested, and decodes 220 tokens/s on an M5 Max. The source says larger models may still suit complex agentic or coding-heavy tasks.

    AIWhy it matters: The source reports benchmark results against several same-tier models and notes where larger models still lead, which helps judge fit for edge agent workloads.

Jul 31

Jul 31Fri
  1. DeepSeek · new models on Hugging FaceAI score75

    DeepSeek releases DeepSeek-V4-Flash-0731 with stronger agentic capabilities

    DeepSeek has released DeepSeek-V4-Flash-0731 as the official version superseding the preview, with substantially enhanced agentic capabilities. The source reports it outperforms DeepSeek-V4-Pro (Preview) on listed benchmarks, including Terminal Bench 2.1 at 82.7 versus 72.1, despite a far smaller activated parameter count. The model ships under the MIT License with DSpark speculative decoding supported in vLLM and SGLang.

    AIWhy it matters: The release shows benchmark gains over the preview and a concrete vLLM and SGLang serving path, useful for teams weighing a self-hosted agentic coding model.

  2. DeepSeek API NewsAI score67

    DeepSeek-V4-Flash API enters public beta with stronger agent benchmarks

    DeepSeek has released the DeepSeek-V4-Flash API in public beta, and developers can use the latest version by setting the model name to deepseek-v4-flash. The source reports agent benchmark results far above V4-Pro-Preview, including 82.7 on Terminal Bench 2.1 and 70.3 on Toolathlon verified. V4-Flash natively supports the Responses API format and is adapted for Codex, while V4-Pro and the APP/WEB models are unchanged.

    AIWhy it matters: The release lists agent benchmark results against V4-Pro-Preview and notes Responses API support for Codex, which helps developers gauge the upgrade's practical effect on their workflows.

Jul 28

Jul 28Tue
  1. JetBrains AI BlogAI score60

    Ponytail Skill Cuts Claude Code Costs 10% But Not the Advertised 54%

    JetBrains tested the ponytail skill for Claude Code across 80 paired tasks and found a median 10.3% cost reduction, with p=0.004. Code written fell about 15% median versus the advertised 54%, reaching 31% on larger builds and little on already-lean tasks. No quality difference was detected, and the skill only self-activated when its ruleset was injected by a plugin hook.

    AIWhy it matters: The benchmark separates advertised savings from measured results and shows the code cut depends on how much the baseline agent over-builds.

Jul 26

Jul 26Sun
  1. Fireworks AI BlogAI score60

    Fireworks AI adds open-weight Kimi K3 with US-only serverless endpoints

    Fireworks AI made the open-weight Kimi K3 available for inference and training on its platform, with US-only serverless endpoints and Zero Data Retention. In its own head-to-head with Opus 5, the post reports K3 at 92.7% accuracy and $0.52 per task on SWE (480) against Opus 5's 94.8% and $1.05, with the vendor claiming up to 5x better cost efficiency per task.

    AIWhy it matters: The post compares Kimi K3 with Opus 5 on accuracy and cost per task, giving readers concrete figures to judge the open model against closed alternatives for their own workloads.

Jul 24

Jul 24Fri
  1. Cat WuAI score66

    Claude Opus 5 released as strong option for long-running autonomous work

    Anthropic introduces Claude Opus 5 as a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price, according to the quoted announcement. The author, who works on the product, says Claude Opus 5 is great at long-running autonomous work and invites users to try it and share feedback.

    AIWhy it matters: The post pairs a new model's long-running autonomous strength with a pricing claim, letting readers weigh capability against cost for agentic workloads.

Jul 23

Jul 23Thu
  1. BAAI · new models on Hugging FaceAI score62

    BAAI releases AREX-Base, a 122B deep research agent model

    BAAI has released AREX-Base, a 122B-total, 10B-activated Mixture-of-Experts deep research agent built on Qwen3.5-122B-A10B with a 262,144-token context. The model uses an inner research loop and an outer self-improvement loop, and the source reports it scoring 82.5 on BrowseComp and 85.4 on GAIA, under Apache 2.0.

    AIWhy it matters: The release pairs a 122B-parameter deep research agent with benchmark tables against frontier and open models, letting readers compare its search-agent results directly.

Jul 21

Jul 21Tue
  1. koray kavukcuogluAI score72

    Google releases Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber

    Google introduces Gemini 3.6 Flash as its workhorse model, with better coding, knowledge work, and multimodal performance while reducing token usage. It also launches Gemini 3.5 Flash-Lite, described as the fastest and most cost-effective 3.5-class model for high-throughput applications, and 3.5 Flash Cyber, a version of 3.5 Flash fine-tuned to find and fix cybersecurity vulnerabilities.

    AIWhy it matters: The post lists three distinct models, each aimed at a different job, so readers can map which one fits coding, high-volume, or security workloads.

  2. JetBrains AI BlogAI score62

    JetBrains Context adds repository indexing to coding agents in early access

    JetBrains has launched JetBrains Context in early access, a repository intelligence layer that builds a semantic index so coding agents can retrieve relevant code without repeated searching. In tests on 205 SWE-bench tasks, 175 production-monorepo tasks, and 1,953 code-localization tasks, it reduced agent turns by up to 68%, latency by up to 59%, and execution cost by up to 48%. It works with Claude Code, Codex CLI, and Junie CLI at no additional cost for JetBrains AI subscribers, and it does not store source code on JetBrains Context servers.

    AIWhy it matters: The source gives benchmark figures for turns, latency, and cost, showing how repository indexing might change agent workflows on large codebases.

Jul 13

Jul 13Mon
  1. Cognition Blog (Devin, Windsurf)AI score62

    Fable 5 with a sidekick costs less than Opus 4.8 on FrontierCode

    Cognition found that Fable 5 led runs cost less than Opus 4.8 led runs on FrontierCode 1.1 when both used the same sidekick, $1.86 versus $2.04 per run. Fable 5 scored 60.7 against 54.6 for Opus 4.8 in those configurations, and it took fewer lead turns, delegated earlier, and rarely edited code itself. The post attributes the difference to delegation style rather than per-token price, and notes that the approach gives little benefit on short or serial debugging tasks.

    AIWhy it matters: The source compares lead-model delegation habits on a coding benchmark, showing how a pricier model can lower total agent cost through fewer turns and better handoffs.

Jul 9

Jul 9Thu
  1. Meta AI BlogAI score72

    Meta releases Muse Spark 1.1 with agent and coding gains

    Meta Superintelligence Labs has introduced Muse Spark 1.1, a multimodal reasoning model aimed at agentic tasks, with gains in tool use, computer use, coding, and multimodal understanding. It supports a 1 million token context window and is available in Thinking mode in the Meta AI app and on meta.ai, with developers able to access it through a public preview of the Meta Model API.

    AIWhy it matters: The post specifies Muse Spark 1.1's agent, coding, and multimodal gains and its Meta Model API preview access, which helps developers judge its fit for their workflows.

Jul 8

Jul 8Wed
  1. Cognition Blog (Devin, Windsurf)AI score62

    Cognition releases SWE-1.7, a coding model trained with long-horizon RL

    Cognition launched SWE-1.7, which it says reaches frontier-level coding performance at lower cost, trained from a Kimi K2.7 base. The post describes RL methods including top-p sampling replay to preserve entropy, compressed weight deltas across multi-cluster training, and self-compaction for rollouts up to six hours. SWE-1.7 is available in Devin via Cerebras at 1000 TPS.

    AIWhy it matters: The post details entropy preservation, multi-cluster weight sync, and self-compaction, offering concrete RL training techniques for long-horizon coding agents to compare against one's own pipeline.

Jul 7

Jul 7Tue
  1. Berkeley AI ResearchAI score62

    Berkeley researchers outline how data systems must change as agents take over knowledge work

    Berkeley AI Research authors argue that near-free inference will make agents the dominant workload for data systems, requiring redesign for agentic speculation, agent-run state and coordination, and agent-synthesized systems. The post cites inference prices falling 9x to 900x per year with a median near 50x, and reports that about 80-90% of sub-queries in a text-to-SQL benchmark were duplicates. It frames the three directions as data systems for, of, and by agents.

    AIWhy it matters: The piece maps three concrete data-system challenges posed by near-free inference, useful for anyone designing infrastructure for agent workloads and memory.

  2. Meta AI BlogAI score75

    Meta launches Muse Image, an agentic image model with search and code tools

    Meta Superintelligence Labs has released Muse Image, which can invoke search and coding tools and self-refine its generations before output. It is available today in the Meta AI app, meta.ai, Instagram Stories in the US, and WhatsApp in limited countries, with Facebook coming soon. Meta also previewed Muse Video, which is coming soon to creators and Meta AI and is reported as ranking No. 3 on Arena for text-to-video at the time of writing.

    AIWhy it matters: The source describes how search, code execution, and self-refinement change image generation, which matters to anyone comparing agentic media models with plain prompt-to-image systems.

Jun 29

Jun 29Mon
  1. Cognition Blog (Devin, Windsurf)AI score62

    Cognition's Devin Fusion routes coding work between two models to cut cost

    Cognition has released a preview of Devin Fusion, a multi-model harness that runs a frontier main agent alongside a cheaper sidekick agent. On FrontierCode 1.1 Extended, the company reports scores near frontier models at up to 60% lower cost per task, and 41% lower cost when paired with Fable 5, which access was suspended from June 12, 2026.

    AIWhy it matters: The post explains a sidekick architecture with cached persistent contexts, which contrasts with advisor-style tools and shows how cost cuts depend on the main model's delegation behavior.