Anthropic finds collaboration breaks down in agent swarm experiments
Testing 45 agents on vulnerability hunting and game-building revealed a pattern where teamwork falls apart
140
Moonshot AI
A Chinese AI developer that makes the chatbot and model "Kimi." It is a regular fixture at the top of open-source model rankings. Headquartered in Beijing, this startup is, despite its name "Moonshot," a software company with no connection to space exploration or aerospace business. Its flagship model Kimi has been released with an emphasis on its ability to handle long context in a single pass, and it is often grouped together with DeepSeek, Alibaba's Qwen, and others as part of the lineage of Chinese-made open-source models. As of August 2026, OpenAI and Anthropic have China AI
Current rank (1M)
31
- Change
- Prev.
- 21
Rank over the last 30 days
2026.08.17 – 2026.09.15 · High 21 · Low 31 · Now 31
Moonshot AI · Related stories
Even the best AI models can't score above 60% on pure vision
Moonshot AI's new benchmark ranks GPT-5.6 Sol first at 59.7%, revealing that many "reasoning errors" are actually perception failures
20
OpenAI, Anthropic slash prices as Chinese AI closes the gap
GPT-5.6 Luna cut 80%, Claude Opus 5 launches at half the price of top-tier model
140
DeepSeek-V4-Pro launches officially with three-tier reasoning intensity control
Released via Expert Mode, V4-Pro edges out Opus-4.8 on Terminal Bench
20
NVIDIA's Nemotron 4 aims for 1 trillion parameters, still trails China
NVIDIA's open-weight Nemotron 4 is said to reach 1 trillion parameters, but Kimi K3 and DeepSeek V4 are already bigger
50
Grok 4.6 unveiled: top-tier performance at half the price
Scoring 61 on the intelligence index, it ties GPT-5.6 Sol and closes to within two points of Claude Opus 5 — yet costs less than half as much. A leap achieved not by scaling up but by changing the training recipe alone. The price war sparked by xAI has only just begun.
20
DeepSeek V4 Pro GA Benchmarks Leak, Nears Top Open-Source Tier
Matches Kimi-K3 and GLM-5.2 in agentic evaluations like Terminal Bench and HLE
10
Grok 4.6 scores 61 on intelligence index, matches GPT-5.6 at the same price
Score jumps 5 points a month after launch, takes the runner-up spot to Claude Opus 5 in agentic tasks
00
AA-AnalystAgent Benchmark: Claude Opus 5 Leads at 54%
Artificial Analysis tests AI on real-world spreadsheets and documents. Claude Opus 5 scored 54%; 57% of failures anchored on a wrong early hypothesis.
20
Anthropic IPO: investors grill company on Chinese AI first
Anthropic is pushing for a September-October listing, but Chinese rival models and data center backlash have emerged as key variables
00
GPT-5.6 model pricing on Bedrock cut by up to 80%
Amazon Bedrock has adjusted on-demand inference pricing for its OpenAI GPT-5.6 model lineup
00
Kimi K3 Inference Benchmarks: Together AI Ranks Among Top Providers
Moonshot AI tested Kimi K3 across major inference providers, with Together AI ranking first or tied for first in 3 of 4 benchmarks
00