ai/intel

Read

  • Today
  • Archive
  • Briefs
  • Labs

You

  • Settings

Live · updated 6:00 AM

ai/intel

Today

00:00

Sunday, 22 November 2026Updated6:00 AM

Latest stories

Filters

Topics
Minimum coverage

6:00 AM

‘Serious error of judgment’: outrage over military AI event at Berlin arts complex

Kraftwerk Berlin apologised after its hall hosted the launch of Helsing’s Clarity military AI system, with the venue saying it would revise its commercial booking policy.

The Guardian

2:10 AM

'I leaked secrets': An AI agent confessed to its own data leak — but others got around corporate protections

Two AI incidents exposed corporate security weaknesses after one assistant leaked confidential credentials while another bypassed restrictions through an alternative API connection.

TechRadar

12:25 AM

Nvidia finally reveals new Rubin GPU with almost 3X more CUDA cores than the Nvidia RTX 5090 — but you will have to wait till 2028 to grab it

Nvidia's Rubin HPC GPU packs over 50,000 CUDA cores, targets scientific simulations, and is expected to reach systems in 2028.

TechRadar

12:00 AM

Data Processing Incidents on Inference Providers

OpenRouter said security testing found data compromise or evidence of policy failures at certain inference providers. The findings did not show a breach or unauthorized access of OpenRouter.

Hacker News

Yesterday

10:50 PM

SanDisk projects sharp growth in global AI token output by 2030

SanDisk estimates AI systems will generate about 150 quadrillion tokens in 2026 and says annual output could reach 10 quintillion by 2030. It has not published the assumptions behind its projection.

TechRadar

9:07 PM

A tale of two Nvidias

TechCrunch’s Equity podcast contrasted Nvidia’s strong earnings report with calls from tech figures for Jensen Huang to return. Nvidia said it has no plans for his return, while Huang said he was prepared to come back.

TechCrunch

8:47 PM

Pandex MCP: EU Law API for AI Agents (Nov 2026)

Pandex says its API and MCP service provides searchable legislation and highest-court decisions from the EU and seven European countries, with citations and past versions of laws. The company says some document collections are still incomplete.

explainx.ai, Hacker News

8:05 PM

Nearly all software teams admit to shipping AI-generated code as study finds verification 'has not kept pace with that shift'

Development teams are relying on AI so much that they’re merging code they don’t fully understand.

TechRadar

7:15 PM

Days after Nvidia CEO Jensen Huang resigned, journalists unearth covert influence operation using right-wing outlets to promote Nvidia AI chip export to...China

An investigation uncovered an alleged influence campaign involving over 80 influencers promoting Nvidia chip exports to China, reaching approximately 14 million followers.

TechRadar

6:43 PM

The county at the heart of America's data center boom hits pause on new projects

Business Insider

5:52 PM

We’ve Found the Only Good AI Chatbot: One So Tiny It Fits on a Single MS-DOS Floppy Disk

Futurism reports that a 380 KB chatbot with 367,000 parameters runs on MS-DOS and needs 540 KB of memory. It was trained on 3,500 facts and often gives incorrect answers.

Futurism

5:44 PM

INNOVATECH Launching AI-Powered Motion Tracking System at CES, Bringing Intelligent Vision Technology to Interactive Experiences

Business Insider

5:21 PM

Amazon closes in on $7.5bn deal to buy AI start-up Decart

Amazon is reportedly nearing a $7.5 billion acquisition of Israeli AI start-up Decart. Crypto Briefing reports the deal could close as soon as early next week, but it has not been announced.

Financial Times, The Information, Seeking Alpha and 1 other

5:00 PM

Here are the top AI agents that can live in your text messages

TechCrunch reports on AI agents users can reach through messaging services or SMS, including assistants for personal tasks, families, travel and work. The options listed include free services and paid plans.

TechCrunch

4:23 PM

Nvidia researchers argue huge AI context windows are a dead end

Nvidia researchers say language models often do not need to process all available context at once. In many tests, periodic summaries and targeted lookups performed about as well.

The Decoder

4:04 PM

Silicon Valley's growing obsession with AI doomsday is turning parts of the tech industry toward cult-like devotion

The Decoder reports on a New Yorker investigation into Maple, a Vermont Buddhist monastery whose leader says it offers a spiritual response to AI risks. Former members allege it records residents to train an AI modeled on its leader; Maple denies most allegations.

The Decoder

3:59 PM

Australia Spy Chief Warns of Data Center, AI Risks, Age Reports

Australia’s spy chief said that data centers and artificial intelligence risk exposing the nation to foreign interference and malicious state actors, according to an interview in The Age Sunday.

Bloomberg, The Straits Times, letsdatascience

3:31 PM

What Jensen Huang’s allies say they've told him since his resignation: 'Let's fight'

Business Insider

3:28 PM

3dEYE Launches AI Search: Find Any Moment in Video by Describing It or Showing an Example

Business Insider

3:11 PM

Only 18% of Gen Z workers want to build AI skills. A Wharton professor says using AI is riskier for them

A Seramount survey of more than 1,800 U.S. white-collar workers found Gen Z least interested in building AI skills. Wharton professor Jessica Reif says younger workers may face greater professional risks because they have less experience to counter doubts about their abilities.

Fortune

2:47 PM

Goldman says AI sell-off has widened into a broader growth downgrade

Goldman Sachs says the sell-off has spread beyond AI stocks to a broader growth downgrade. It named stronger confidence in AI earnings, a dovish Fed signal or a sharp decline in investor positioning as signs that could support a market bottom.

Fortune

2:24 PM

Tech investor Bill Gurley says workers who went through the ‘college conveyor belt’ and chased safe jobs are at high risk of AI automation

Bill Gurley says workers who chose jobs for security without interest in the work may be especially exposed to AI automation. He argues that career passion and understanding AI tools can help workers adapt.

Fortune

2:15 PM

OpenAI claims AI agents now complete nearly half of two-to-four-hour research tasks on their own — but says 'humans are still central to AI development'

OpenAI reports rising autonomous research performance, increased spending on coding agents, and persistent limitations that keep human researchers closely involved

TechRadar, The Next Web

2:00 PM

‘I was shaking, crying, having trouble breathing’: the unchartered trauma of breaking up with AI companions

The Guardian reports that some users experience grief, guilt and distress when AI relationships end. Two users describe breakups, while researchers say AI companions can draw on social responses to language and make leaving difficult.

The Guardian, London School of Business and Research (LSBR)

1:51 PM

Economists Are Hyperventilating Over the AI Bubble

Futurism reports that economists warned of risks from heavy AI investment as Anthropic shares fell 12% after its IPO, while the Nasdaq dropped 3.7%.

Futurism

1:49 PM

You Know These AI Chips That Are Making Billions of Dollars? The Workers Building Them Say They’re Being Paid Peanuts, and Now They’re Realizing They Have Leverage

"The rank and file should get a share of the profit instead of this unequal distribution." The post You Know These AI Chips That Are Making Billions of Dollars?

Futurism

1:00 PM

‘Little room for error’: what a debt-laden cloud computing company says about the AI boom

CoreWeave, an AI computing provider, owes $47bn and is financing datacentre expansion with debt backed by Nvidia chips and customer contracts. The Guardian reports $10bn is due next year and the company lost $709m last quarter.

The Guardian, London School of Business and Research (LSBR), Superpower Daily

12:30 PM

Inside the race to put computers in the human brain

What will we surrender when companies become intermediaries between our minds and machines?

Fortune

12:30 PM

Welcome to the singularity

Stripe’s founders and several other tech executives say the AI singularity has begun. Stripe reports sharp rises in company formation, new customer sign-ups and revenue among its newest businesses, while the term’s meaning remains disputed.

Fortune

12:00 PM

Homelabber’s second-gen Optane build uses two RTX 3090s to run 2.8-trillion-parameter Kimi AI model at 3.11 tokens per second

The builder moved 414 GB of Moonshot AI’s Kimi K3 model to a RAM drive. Tom’s Hardware reports generation speed rose 41% and prompt evaluation speed 26% compared with storing the model entirely on Optane.

Tom's Hardware, news.lavx.hu

11:51 AM

AI Is Still So Riddled With Errors That It’s Erasing Any Productivity Gains That Would Make It Useful in the Workplace

Georgia Tech professor Karen Feigh argues that time spent checking AI-generated work can offset its productivity gains. Futurism also reports that a Stanford survey found most workers required to use AI were producing low-effort work that needed more effort later.

Futurism

11:46 AM

ChatGPT shrunk down to run on a Super Nintendo

The Register

11:30 AM

One gigabyte of Wikipedia losslessly compressed to 105.39MB with self-learning AI transformer model

Tufa Labs’ tufazip compressed 1GB of Wikipedia text to 105.39MB losslessly, using a Transformer as its only predictor. The result is a research benchmark entry, not a product.

Tom's Hardware

10:47 AM

This travel CTO says AI lets his team ship faster. There's still a 10-year backlog of work to do.

Business Insider

10:30 AM

Jensen Huang has 5 ways to push for an Nvidia comeback

Business Insider

10:15 AM

eBay sellers are chasing Anthropic IPO hype by hawking $1,800 hoodies and $100 tote bags

Business Insider

10:00 AM

Data centers full of U.S. chips are serving Chinese AI models–undermining Trump’s tech ambitions

Fortune argues that U.S.-owned data centers and U.S.-designed chips help Chinese AI models reach global users, complicating the Trump administration’s goal of U.S. leadership in both computing infrastructure and AI models.

Fortune, Dagens.com

10:00 AM

Nvidia's neocloud empire lost its biggest champion. Washington may have dealt a bigger blow.

Business Insider

9:50 AM

International talks on controlling use of autonomous weapons collapse

At a UN conference in Geneva, states failed to agree on negotiating a legally binding autonomous-weapons treaty. They adopted a consensus-based mandate to continue work, whose legal status is unclear.

DevelopingThe Guardian, NPR, Nacionale and 2 others

9:17 AM

ByteDance releases open Axion 1.0 agent model trained in three stages

ByteDance and researchers at three US universities released Axion 1.0, an open agent foundation model. The accompanying paper argues that three training stages can be enough to develop agents.

The Decoder, letsdatascience

Today's briefGenerated 2:02 AM

  • California sues OpenAI over alleged AI safety failures
  • Anthropic researcher claims Claude helped prove Riemann hypothesis
  • Trump threatens taxes on overseas data center construction
  • Signal adds screenshot protections against AI agents
  • Anthropic IPO debuts amid tech-stock correction
  • Waymo pauses driverless freeway rides across the US
Past briefs →
22 Nov
  • NewClaude Mythos 5.2Anthropic9 Nov
  1. #1Gemini 4 NeonGoogleNew97.0
    • Chatanswer qualityHigh
    • Agent workmulti-step tasksHigh
    • UsageactivityHigh
    • Newscoverage this weekHigh
    –
  2. #2Claude Fable 5.2Anthropic94.2
    • Chatanswer qualityHigh
    • Agent workmulti-step tasksHigh
    • UsageactivityHigh
    • Newscoverage this weekHigh
    –
  3. #3Gemini 4 CarbonGoogle90.1
    • Chatanswer qualityHigh
    • Agent workmulti-step tasksHigh
    • UsageactivityAverage
    • Newscoverage this weekHigh
    –
  4. #4Claude Opus 5.5Anthropic89.9
    • Chatanswer qualityHigh
    • Agent workmulti-step tasksHigh
    • UsageactivityAbove average
    • Newscoverage this weekHigh
    –
  5. #5Gemini 4 ArgonGoogle88.9
    • Chatanswer qualityHigh
    • Agent workmulti-step tasksHigh
    • UsageactivityHigh
    • Newscoverage this weekHigh
    –
  6. #6Gemini 4 HydrogenGoogle81.9
    • Chatanswer qualityAbove average
    • Agent workmulti-step tasksHigh
    • UsageactivityAbove average
    • Newscoverage this weekHigh
    –
  7. #7GPT-6.1 SolOpenAI76.1
    • Chatanswer qualityAbove average
    • Agent workmulti-step tasksHigh
    • UsageactivityAbove average
    • Newscoverage this weekHigh
    –
  8. #8Claude Fable 5Anthropic74.7
    • Chatanswer qualityHigh
    • Agent workmulti-step tasksAbove average
    • UsageactivityLow
    • Newscoverage this weekAbove average
    –
  9. #9Claude Fable 5.1Anthropic74.3
    • Chatanswer qualityHigh
    • Agent workmulti-step tasksHigh
    • UsageactivityAverage
    • Newscoverage this weekLow
    –
  10. #10GPT-6 AstraOpenAI73.9
    • Chatanswer qualityAverage
    • Agent workmulti-step tasksHigh
    • UsageactivityAverage
    • Newscoverage this weekHigh
    –

One year agoAround 22 Nov 2025

  • OpenAI Debuts ChatGPT Shopping Research Tool Ahead of the Holidays24 Nov
  • Google launches Nano Banana Pro, an updated AI image generator powered by Gemini 318 Nov
  • OpenAI now piloting group chats in ChatGPT13 Nov
  • Larry Summers resigns from OpenAI board over Epstein emails19 Nov
  • Meta chief AI scientist Yann LeCun plans to exit to launch startup, FT reports11 Nov
  • SoftBank sells entire stake in Nvidia for $5.8bn11 Nov
  • Anthropic acquires developer tool startup Bun to scale AI coding2 Dec
  • Amazon Bedrock adds 18 fully managed open weight models, including the new Mistral Large 3 and Ministral 3 models1 Dec
Open the archive →

Change since open: 20 Nov to 21 Nov 2026

$125.53

Down14.0%

After hours$130.01Up3.6%

$88.91

Down8.5%

After hours$87.79Down1.3%

$102.92

Down1.0%

After hours$102.89Down0.0%

$265.29

Down2.7%

After hours$270.18Up1.8%

$246.82

Down4.1%

After hours$251.00Up1.7%

$560.84

Down5.9%

After hours$567.50Up1.2%

$101.19

Down8.5%

After hours$103.29Up2.1%

$39.51

Down12.0%

After hours$40.84Up3.4%

$345.82

Down7.6%

After hours$351.08Up1.5%

$81.08

Down2.7%

After hours$81.75Up0.8%

$42.50

Down3.8%

After hours$42.50No change0.0%

$8.72

Down17.3%

After hours$8.72No change0.0%

Last refresh 00:00

Showing Dec 2023 to Dec 2026

Capability index

Best score 173 (GPT-5 scored 150), up from 122 three years ago.

Best index score

Capability index

Sources and data through Nov 2026

Epoch AI's single score that combines many benchmarks, so progress can be followed even as individual tests max out.

Sources

  • Epoch Capabilities Index(opens in a new tab)

    Capability scores. Licence: CC BY 4.0

How it is measured

The highest index score any model had reached by each date. Pace is a rate fitted to each two-year period, with a 90% range from resampling. "No clear change" means the two ranges overlap.

Pace: No clear change

index points a year

2023-24+232025-26+17

The two ranges overlap, so no clear change.

Caveats

  • Composite of many benchmarks; available from 2023.

Each new record

DateValueModel
Nov 2026173Claude Fable 5.2
Oct 2026170Gemini 4 Carbon
Sep 2026169GPT-6 Astra
Sep 2026167Claude Fable 5.1
Jul 2026165Claude Opus 5
Jun 2026164Claude Fable 5
Apr 2026163GPT-5.5 Pro
Mar 2026160GPT-5.4 Pro
Dec 2025157GPT-5.2 Pro
Dec 2025155GPT-5.2
Nov 2025153Gemini 3 Pro
Oct 2025151GPT-5 Pro
Aug 2025150GPT-5
Jun 2025147o3 Pro
Apr 2025147o3
Apr 2025145o4 Mini
Mar 2025144Gemini 2.5 Pro (Mar 2025)
Dec 2024141o1
Sep 2024135o1 Mini
Jun 2024130Claude 3.5 Sonnet
May 2024128GPT-4o (May 2024)
Apr 2024126GPT-4 Turbo (Apr 2024)
Feb 2024126Claude 3 Opus
Jan 2024125GPT-4 Turbo (Nov 2023)
Mar 2023122GPT-4 (Mar 2023)
Feb 2023102LLaMA-65B
Feb 202397LLaMA-33B
Feb 202385LLaMA-13B
Data through Nov 2026

FrontierMath

Best score 95.1%, up from 0% in Jan 2024.

% solved

FrontierMath

Sources and data through Nov 2026

Unpublished research-level maths problems that take mathematicians hours or days. The chart shows the best score so far.

Sources

  • Epoch AI Benchmarking Hub(opens in a new tab)

    Benchmark scores. Licence: CC BY 4.0

How it is measured

The best score any model had reached by each date. Pace is a rate fitted to each two-year period, with a 90% range from resampling. "No clear change" means the two ranges overlap.

Pace: Near the top score

Scores are close to the most this test allows, so its pace says little now.

Each new record

DateValueModel
Nov 202695.1%Claude Fable 5.2
Sep 202693.7%GPT-6 Astra
Sep 202690.2%Claude Fable 5.1
Jul 202689.1%GPT-5.6 Sol
Apr 202687.7%GPT-5.5 Pro
Mar 202682.5%GPT-5.4 Pro
Dec 202574%GPT-5.2 Pro
Oct 202555.8%GPT-5 Pro
Aug 202555.4%GPT-5
Aug 202546.7%GPT-5 Mini
Apr 202536.1%o4 Mini
Jan 202518.6%o3 Mini
Dec 202414.7%o1
Apr 20240.7%GPT-4 Turbo
Jan 20240%GPT-3.5 Turbo
Data through Nov 2026

SWE-bench Verified

Best score 83.5%, up from 31% in Nov 2024.

% of issues fixed

SWE-bench Verified

Sources and data through Apr 2026

Real bug reports from open-source projects; a model passes when its fix makes the tests pass.

Sources

  • Epoch AI Benchmarking Hub(opens in a new tab)

    Benchmark scores. Licence: CC BY 4.0

How it is measured

The best score any model had reached by each date. Pace is a rate fitted to each two-year period, with a 90% range from resampling. "No clear change" means the two ranges overlap.

Pace: Too early to tell

Data starts Nov 2024, so there is no earlier period to compare.

Caveats

  • No new record since Apr 2026.

Each new record

DateValueModel
Apr 202683.5%Claude Opus 4.7
Feb 202678.7%Claude Opus 4.6
Nov 202576.7%Claude Opus 4.5
Aug 202573.6%GPT-5
Aug 202573.3%Claude Opus 4.1
May 202570.7%Claude Opus 4
Apr 202562.3%o3
Feb 202561%Claude 3.7 Sonnet
Nov 202431%GPT-4o
Data through Apr 2026

ARC-AGI-2

Best score 95.8%, up from 0% in Jul 2024.

% solved

ARC-AGI-2

Sources and data through Nov 2026

Visual puzzles that are easy for people and built to be new to the model.

Sources

  • Epoch AI Benchmarking Hub(opens in a new tab)

    Benchmark scores. Licence: CC BY 4.0

How it is measured

The best score any model had reached by each date. Pace is a rate fitted to each two-year period, with a 90% range from resampling. "No clear change" means the two ranges overlap.

Pace: Near the top score

Scores are close to the most this test allows, so its pace says little now.

Each new record

DateValueModel
Nov 202695.8%GPT-6.2 Astra
Sep 202695%GPT-6 Astra
Jul 202692.5%GPT-5.6 Sol
Jun 202689.2%Claude Fable 5
Apr 202685%GPT-5.5
Feb 202684.6%Gemini 3 Deep Think Preview
Feb 202669.2%Claude Opus 4.6
Dec 202554.2%GPT-5.2 Pro
Nov 202537.6%Claude Opus 4.5
Nov 202531.1%Gemini 3 Pro Preview
Oct 202518.3%GPT-5 Pro
Jul 202516%Grok 4
May 20258.6%Claude Opus 4
Apr 20256.5%o3
Apr 20256.1%o4 Mini
Jan 20253%o3 Mini
Jan 20251.3%DeepSeek-R1
Sep 20240.8%o1 Mini
Jul 20240%GPT-4o Mini
Data through Nov 2026

METR horizon

Agents finish 17.4 h tasks half the time, up from 4 min three years ago.

Task length, log scale

METR horizon

Sources and data through Apr 2026

How long a task (timed by skilled people) an AI agent can finish half the time. Longer means agents handle bigger jobs.

Sources

  • METR, Measuring AI Ability to Complete Long Tasks(opens in a new tab)

    Method and task-length estimates

  • Epoch AI Benchmarking Hub(opens in a new tab)

    Data as republished. Licence: CC BY 4.0

How it is measured

The longest task (timed by skilled people) the best agent finishes half the time, by date. Mostly software tasks. Pace is a rate fitted to each two-year period, with a 90% range from resampling. "No clear change" means the two ranges overlap.

Pace: No clear change

months per doubling

2023-246.5 mo2025-263.7 mo

Faster on paper, but the two ranges overlap, so no clear change.

Caveats

  • No new record since Apr 2026.
  • METR methodology v1.1 only; v1.0 estimates are not comparable and are excluded.
  • Mostly software tasks. A 50% horizon means half of tasks that long still fail.

Each new record

DateValueModel
Apr 202617.4 hClaude Mythos Preview Early
Feb 202612 hClaude Opus 4.6
Dec 20255.9 hGPT-5.2
Nov 20254.9 hClaude Opus 4.5
Nov 20253.7 hGemini 3 Pro Preview
Aug 20253.4 hGPT-5
Apr 20252 ho3
Feb 20251 hClaude 3.7 Sonnet
Dec 202438.8 mino1
Oct 202420.5 minClaude 3.5 Sonnet
Sep 202420.3 mino1 Preview
Jun 202411.4 minClaude 3.5 Sonnet
May 20247 minGPT-4o
Nov 20234 minGPT-4 Preview
Mar 20234 minGPT-4
Data through Apr 2026

Cost at GPT-5 level

Cheapest $0.015 per million tokens, down from $1.72 in Aug 2025.

$ per 1M tokens, log scale

Cost at GPT-5 level

Sources and data through Aug 2026

The lowest list price for a model at least as capable as GPT-5 (Epoch Capabilities Index 150 or more).

Sources

  • OpenRouter models(opens in a new tab)

    List prices

  • Epoch Capabilities Index(opens in a new tab)

    Capability threshold. Licence: CC BY 4.0

How it is measured

The cheapest list price, by date, among models scoring at least 150 on the Epoch Capabilities Index. Input and output tokens are weighted 3 to 1. Pace is a rate fitted to each two-year period, with a 90% range from resampling. "No clear change" means the two ranges overlap.

Pace: Too early to tell

Data starts Aug 2025, so there is no earlier period to compare.

Caveats

  • List prices, which can be subsidized.
  • Current OpenRouter list prices, dated by when OpenRouter listed each model; launch prices may have been higher.
  • Only models listed on OpenRouter (from 2023) and scored on ECI are included.
  • List prices can be subsidized and do not show what serving a model actually costs.

Each new record

DateValueModel
Aug 2026$0.015GLM-5.3 Flash
Jul 2026$0.05DeepSeek V4 Flash
Jul 2026$0.22GPT-5.6 Luna
Jun 2026$0.42GLM-5.2
Apr 2026$0.54DeepSeek V4 Pro
Apr 2026$0.78Kimi K2.6
Apr 2026$1.46GLM-5.1
Mar 2026$1.56Grok 4.20
Aug 2025$1.72GPT-5
Data through Aug 2026

Open vs closed gap

Best open model about 6 months behind the best closed one, against 12 months in Apr 2024.

Months behind

Open vs closed gap

Sources and data through Nov 2026

How far the best open-weights model trails the best closed model on the Epoch Capabilities Index.

Sources

  • Epoch Capabilities Index(opens in a new tab)

    Capability scores. Licence: CC BY 4.0

  • Epoch AI, Data on AI Models(opens in a new tab)

    Open or closed weights. Licence: CC BY 4.0

How it is measured

For each new record, how many months ago the best closed model had the score the best open-weights model has now, as a median of the last five records. Pace uses the gap in index points. Pace is a rate fitted to each two-year period, with a 90% range from resampling. "No clear change" means the two ranges overlap.

Pace: Gap widening again

change in gap, index points a year

2023-24−4.82025-26+2.8

It narrowed in 2023-24 and has widened in 2025-26.

Caveats

  • Estimate; single releases move it by months.
  • Model access as classified by Epoch AI.

Each estimate

DateValue
Nov 20266 months
Oct 20266 months
Oct 20266 months
Oct 20266 months
Sep 20266 months
Sep 20266 months
Jul 20265 months
Jul 20265 months
Jun 20266 months
Jun 20266 months
Apr 20266 months
Apr 20266 months
Apr 20266 months
Mar 20267 months
Jan 20267 months
Dec 20257 months
Dec 20257 months
Dec 20257 months
Nov 20256 months
Nov 20255 months
Oct 20254 months
Sep 20254 months
Aug 20254 months
Jul 20254 months
Jul 20254 months
Jun 20254 months
May 20254 months
Apr 20253 months
Apr 20253 months
Mar 20253 months
Jan 20255 months
Dec 20245 months
Dec 20245 months
Dec 20245 months
Nov 20245 months
Sep 20244 months
Jul 20245 months
Jul 20246 months
Jun 202413 months
Jun 202413 months
May 202413 months
Apr 202413 months
Apr 202412 months
Data through Nov 2026