Only the news that moves the race.
No hype cycle, no rumour mill. Each story is rated for impact and comes with a plain explanation of why it matters.
October 2026
- Oct 17Model releases
DeepSeek ships V4.1 Flash under an MIT licence
DeepSeek released V4.1 Flash, a fast, cheap model with built-in search, under the permissive MIT licence, priced at $0.30 / $1.20 per million tokens via its API.
Why it matters
It resets the floor for what near-frontier coding help costs, and the MIT licence lets any company build on it commercially with almost no strings attached.
V4.1 Flash#19V4 Pro#16 - Oct 15Model releases
Meta returns to open weights with Muse Spark Open
Meta Superintelligence Labs released Muse Spark Open, an open-weights model, three days after launching its closed flagship Muse Wonder.
Why it matters
After a year of closed releases, Meta is again putting competitive weights in the open, keeping pressure on Chinese labs that have dominated the open leaderboard.
Spark Open#12Muse Wonder#8 - Oct 13Model releases
Google launches Gemini 4 Argon with a 2M-token context
Gemini 4 Argon arrived with a 2-million-token context window, native video understanding and top scores on several multimodal benchmarks.
Why it matters
Argon takes the lead in our multimodal and long-context races, giving Google its strongest position in the overall standings this year and narrowing the gap with Claude Opus 5.5 to a few points.
Gemini 4 Argon#5 - Oct 12Model releases
Meta's Muse Wonder targets creative and social use
Meta launched Muse Wonder, a closed multimodal flagship tuned for creative writing, voice and image understanding across its apps.
Why it matters
It brings a frontier-class assistant to billions of users inside WhatsApp, Instagram and Messenger, the largest distribution of any model in our races.
Muse Wonder#8 - Oct 9Research & AGI
Frontier agents now finish tasks that take people a full work week
New time-horizon measurements put the best agents at roughly 50% reliability on software tasks that take skilled humans around 30 hours.
Why it matters
The length of tasks agents can complete has been doubling roughly every few months. If the trend holds, multi-week autonomous projects arrive within a couple of years.
Fable 5.1#4Opus 5.5#1GPT-6 Astra#3 - Oct 7Model releases
Claude Haiku 5.5 brings near-Sonnet quality to the fast tier
Anthropic released Claude Haiku 5.5, its small, fast model, with strong computer use and coding at a fraction of Sonnet's price.
Why it matters
Fast-tier models power most high-volume products and sub-agents. Haiku 5.5 makes capable agent workers far cheaper to run in parallel.
Haiku 5.5#14Sonnet 5.5#6 - Oct 6Model releases
Mistral previews Large 4 with 160+ languages
Mistral opened a preview of Mistral Large 4, a trillion-parameter model with broad multilingual coverage and EU hosting, with weights promised to developers.
Why it matters
It edges the multilingual race in a photo finish and gives European governments and regulated firms a credible sovereign option.
Large 4#17 - Oct 6Model releases
Nano Banana 2.1 tightens Google's grip on image editing
Google's Nano Banana 2.1 improves multi-image consistency and text rendering, retaking the lead in our image editing race.
Why it matters
Conversational image editing is becoming the default way people create images; consistency across edits is the feature creators ask for most.
Nano Banana 2.1#2GPT Image 2#1 - Oct 2Infrastructure
The price of frontier-level intelligence fell 10x in a year
Analysis of API prices shows the cost of reaching a fixed capability level has dropped roughly tenfold in twelve months, driven by efficient small models and cheaper inference hardware.
Why it matters
Tasks that were too expensive to automate last year are now viable, from summarising every support call to running agents on every pull request.
3.5 Flash#10Haiku 5.5#14V4.1 Flash#19 - Oct 1Policy & safety
Labs back a shared standard for agent security
Major labs and agent platforms published a joint framework for defending agents against prompt injection, covering permissions, provenance and human approval.
Why it matters
Prompt injection is the biggest unsolved risk for agents that browse and act. A shared baseline helps buyers compare products and pushes laggards to catch up.
September 2026
- Sep 30Research & AGI
AI system helps settle an open combinatorics conjecture
Mathematicians credited an AI reasoning system with finding the key construction in a proof of a decades-old combinatorics conjecture, verified in the Lean proof assistant.
Why it matters
It is a clear example of AI contributing original mathematics rather than solving known problems, with machine-checked verification removing doubt about correctness.
- Sep 29Model releases
OpenAI's GPT-6.1 Sol becomes the writer's model
GPT-6.1 Sol, OpenAI's balanced model, shipped with major gains in creative and long-form writing and a lower price than Astra.
Why it matters
It narrowly takes the lead in our creative writing race, a category Claude had held for most of the year, and gives OpenAI a strong workhorse below Astra.
GPT-6.1 Sol#2GPT-6 Astra#3 - Sep 25Use cases
Coding agents now author most pull requests at surveyed tech firms
A survey of engineering organisations found that a majority of merged pull requests now originate from coding agents, with humans reviewing and steering.
Why it matters
Software engineering is the first profession where agentic AI has become the default workflow, a preview of what comes next for other knowledge work.
Opus 5.5#1Sonnet 5.5#6GPT-6 Astra#3 - Sep 22Model releases
Claude Opus 5.5 retakes the overall lead
Anthropic's Claude Opus 5.5 launched with the strongest agentic coding results yet recorded and moved to the top of our overall race.
Why it matters
Coding and agents are where AI spend is concentrating. Opus 5.5 leading both makes it the default choice for many engineering teams, though its overall lead is a photo finish.
Opus 5.5#1GPT-6 Astra#3 - Sep 21Model releases
xAI's Grok 4.7 doubles down on speed and real-time data
Grok 4.7 improves reasoning and math while keeping live access to X data for real-time answers.
Why it matters
xAI moves back into the overall top ten, and real-time data remains a differentiator for news and market research.
Grok 4.7#7 - Sep 15Infrastructure
Gigawatt-scale training clusters come online
Several labs and cloud providers brought training campuses of a gigawatt or more online this quarter, each housing hundreds of thousands of accelerators.
Why it matters
Compute is the raw input for the next generation of models. These sites set the scale of training runs expected in 2027.
- Sep 10Robotics
Humanoid robots move into paid warehouse work
Several humanoid robot makers moved from demos to paid deployments moving totes and loading trucks in logistics warehouses.
Why it matters
Vision-language-action models are letting robots generalise across tasks. Paid deployments are the first sign the economics can work.
- Sep 3Model releases
GPT-6 Astra arrives with leading computer use and math
OpenAI's GPT-6 Astra launched as its new flagship, with state-of-the-art computer use and strong agentic coding.
Why it matters
Astra took the overall lead for three weeks and still leads our computer-use race. Agents that operate real software are the next big platform shift.
GPT-6 Astra#3
August 2026
- Aug 28Use cases
AI-designed antibiotic candidate enters clinical trials
A novel antibiotic designed with generative models against drug-resistant bacteria began phase 1 human trials.
Why it matters
Drug discovery timelines are long, so early AI-designed candidates reaching trials are an important test of whether AI shortens them.
- Aug 20Model releases
Claude Fable 5.1 pushes deep reasoning further
Anthropic's Fable 5.1, its slow, deliberate reasoning model, set records on frontier reasoning benchmarks and very long agentic tasks.
Why it matters
Fable leads our hardest reasoning races. For research and analysis where accuracy matters more than speed, it is the model to beat.
Fable 5.1#4 - Aug 2Policy & safety
EU begins enforcing AI Act rules for general-purpose models
The EU AI Office gained enforcement powers over general-purpose AI model obligations, including documentation, copyright policies and risk assessment for the most capable models.
Why it matters
Every major lab serving European users must now show regulators how its models are trained and tested, shaping disclosure norms worldwide.
July 2026
Stories are illustrative placeholders modelled on real patterns of AI news, written to demonstrate the format. Ranks shown are current standings in each model's primary race.