Search “Agents”

Google Releases Gemini 4 Carbon Model, Focused on Long-Horizon Agentic Workflows

Google has released its flagship model Gemini 4 Carbon, built specifically for continuously running agentic workflows, with a focus on improving coding, agentic, scientific task, and full-stack web development capabilities. The model ranks first on the Artificial Analysis Intelligence Index v5.0 with a score of 59, surpassing Claude Opus 5.5, GPT-6 Astra, and GPT-6.1 Sol; the launch promotional price is $2 per million input tokens and $10 per million output tokens, with a 95% discount available for cached input tokens. Carbon adopts a phased rollout strategy, first opening to select cybersecurity defenders, then expanding the range of trusted partners through the Fairwind program, and afterward opening to developers, enterprises, and consumers.

Google Releases Gemini 4 Carbon, Strengthening Cybersecurity and Long-Running Agent Workflows

Two weeks after the debut of Gemini 4 Argon, Google released a new-generation frontier model, Gemini 4 Carbon, saying it achieves the current highest level in cybersecurity and long-horizon tasks, with enhancements for coding, Agent workflows, and cyber defense. The model scored 88.2% on the vulnerability discovery benchmark SEC-bench Pro, 54.8% on the exploit benchmark Exploit Gym, and 71.0% on the patch application benchmark CWE-bench. Carbon will first be made available this week to some cybersecurity defenders, then expanded to more trusted partners, and only afterward released to developers, enterprises, and consumers; during the trial period it follows Argon pricing at $2 per million input tokens and $10 per million output tokens, adjusted to $4 and $20 after the trial period.

Kimi K3.1 Pro Released: Intelligence Index Ranks Fourth Globally, Agent Benchmark Nears Opus 5.5

Moonshot AI has released Kimi K3.1 Pro, which ranks fourth globally on the Artificial Analysis Intelligence Index with a score of 55, an 11-point improvement over its predecessor. On Terminal-Bench 4.0, it rose from 13% to 65.7%, narrowing the gap with Opus 5.5 to less than 1%. The model has 3 trillion-level parameters and supports a 1 million context window; it was post-trained on K3 with increased reinforcement learning compute, and a Highspeed version with 3x faster output has also been launched. The 3D Make feature is now available on kimi.com and the Kimi app.

Moonshot AI Releases Kimi K3.1 Pro: A 3T-Class Open-Weight Model Strengthening Coding, 3D and Finance Agents

Moonshot AI has released Kimi K3.1 Pro, calling it the strongest 3T-class open-weight model, featuring a 1M-token context and multimodal capabilities. It simultaneously launched a Highspeed version that is about 2–3× faster, with full weights planned to be open-sourced before 2026-11-04. The model scores 65.7% on Terminal-Bench 4.0, above GPT-6 Astra's 58.2% and slightly below Claude Opus 5.5's 66.4%; it scores 92.7% on BenchCAD with tools and 64.4% on OfficeQA Pro v2. The open-weight version removes cyber-related data from both mid-training and RL, and its xdev-bench-hard score drops from 90.5% to 81.0%.

English translation is being prepared.

MiniMax Releases M3.1 Flash, Scoring 29.6% on Agents' Last Exam to Rank First Among Chinese Models

MiniMax has officially released M3.1 Flash, the model previously known as Space Bunny (Yutu) that topped the charts anonymously on OpenRouter. It scores 29.6% on Agents' Last Exam, ranking first among Chinese models. The model achieves 73.7% on DeepSWE v1.1, 82.2 points on CyberGym, and 61.7 points on OSWorld V2, with output speeds of up to 200+ TPS and a P50 maintained at 170+ TPS. Pricing is 1 yuan per million input tokens, 4 yuan for output, and 0.1 yuan for cache reads.

English translation is being prepared.

Reflection AI open-sources Beam-1.0-501B-A23B, ranked 12th on the Agent Arena open-source leaderboard

Reflection AI has released the open-source model Beam-1.0-501B-A23B under the Apache 2.0 license. It ranks 12th among open-weight models on Agent Arena with a Net Improvement of -3.67%, and ranks 22nd on the Code Arena: WebDev open-source leaderboard with a score of 1438. On the Agent Arena leaderboard it is the highest-ranked US open-source model, ahead of Hy3, MiMo V2.5 Pro and MiniMax M3, but its Net Improvement error margin reaches ±3.84%, making it statistically closer to the two Qwen3.8 models above it. Of the top 25 open-source models on Code Arena: WebDev, 24 come from Chinese labs; Beam 1.0 is the only exception and the highest-ranked US open-source model on that leaderboard, slightly ahead of Kimi K2.5 Thinking and GLM-5.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.