Skip to content
All systems operational0 AI providers monitored, polled every 2 minutes
Live status
LIVE/15+ SOURCES/10+ API MONITORS/UPDATED EVERY 10 MIN/74 events/min

The AI Pulse

Real-time news, model tracking, and ecosystem data for the AI industry. Readable by humans. Structured for agents.

Events per min
74
up 14% vs. 24h avg
Services monitored
23/23 online
all systems operational
Last model release
6m ago
Mistral Medium 3
Last check
28s
23 providers

/ StatusLive API status across every major provider

Polled every two minutes. Latency is p95 over the last hour.

Full status page
Claude
OK
230ms
p95 last hour52s ago
OpenAI API
OK
242ms
p95 last hour21s ago
Google Gemini
OK
266ms
p95 last hour52s ago
GitHub Copilot
OK
97ms
p95 last hour20s ago
Cursor
OK
91ms
p95 last hour21s ago
Perplexity
OK
111ms
p95 last hour48s ago
Groq
OK
94ms
p95 last hour38s ago
AWS Bedrock
OK
262ms
p95 last hour33s ago
Azure OpenAI
OK
228ms
p95 last hour56s ago
Hugging Face
OK
263ms
p95 last hour48s ago
Replicate
OK
93ms
p95 last hour15s ago
Cohere
OK
258ms
p95 last hour57s ago

/ OriginalsEditorial from TensorFeed

Opinionated analysis from our editorial team, published multiple times per week.

All originals
EDITORIAL / ANALYSIS

Claude Just Formalized Fermat in 11 Days on 6 Billion Tokens. The Harness Thesis Got Its Receipt.

Anthropic published the writeup on Friday, September 5, 2026: dozens of Claude agents produced the first complete machine-checked Lean 4 formalization of Fermat's Last Theorem. Eleven days of wall clock, roughly 6 billion output tokens, about 13 million lines of Lean (five times Lean's standard mathlib), around 29,500 intermediate theorems in the final graph, and independent verification by Lean itself. The theorem is not the news. Andrew Wiles cleared Fermat in 1994 across 129 journal pages and seven years of solitary work. What is new is the shape of the compute that produced a mechanically checkable version of the same result: a general-purpose model every enterprise customer already has, wrapped in a graph of proof obligations, running in parallel across dozens of agents for eleven days, at a compute bill on the order of a seed check. The harness is the product, and this run is the first public receipt with a number attached to every column. The numbers table (wall clock 11 days launched by Anthropic + Columbia early to mid August 2026, output tokens ~6B across all agents and rewrites and failed branches, Lean lines produced ~13M roughly 5x the size of Lean's standard mathlib, intermediate theorems ~29,500 nodes in the shared proof graph, concurrency dozens of agents working the graph in parallel via Prove2Me, verifier Lean 4 deterministic with the trust root outside the model, all figures from Anthropic's September 5 research post with token figures counting output only). Two things read straight off the table: this run consumed more output tokens than most public model evaluations combined (six billion output tokens on a single research project is a compute footprint in the neighborhood of a small pretraining fine-tune not a benchmark), and the concurrency figure is the interesting one (dozens of agents running for eleven days is not a serial reasoning trace it is a parallel construction which is exactly the shape a proof of this scale requires and exactly the shape a single-agent chatbot cannot express). Why the default harness failed: the important sentence in the writeup is the one about the first attempt (Anthropic tried the standard Claude Code multi-agent workflow the same harness shape thousands of teams run in production every day and it broke, individual agents made local progress then they lost track of the overall project state and stopped coordinating, long-horizon memory degradation is the canonical failure mode of the entire agent stack and it fired on the hardest available test), the fix was not a bigger model a longer context window or a smarter prompt but a different orchestrator (Tianyi Peng an Anthropic researcher who initiated the project built Prove2Me with collaborators at Columbia, Prove2Me maintains a directed acyclic graph of theorem statements, an agent picks an unfinished node drafts a proof against the statement runs Lean and either commits a verified subtree or fails and hands the node back to the queue, the graph is the memory the individual context windows cannot hold and the graph is also what lets a dozen agents work at once without stepping on each other's proof state); that is the whole insight and it is a small one on paper (in practice it is the difference between an agent fleet that quietly stalls after three days and one that converges on a formal proof of the Modularity Theorem in eleven, the model contributed general reasoning and the harness contributed the ability to remember and to parallelize, every commercial agent product in the market is going to have to answer the same design question inside the next year because the shape that worked for Fermat is the shape a large software refactor needs or a multi-week security audit or a data-migration project). What the receipt costs: Anthropic did not publish the compute bill so build it from public pricing (six billion output tokens at the Fable 5 rate of $50 per million is $300,000 in output alone, agentic workloads at this shape run inputs roughly 20 to 50 times higher than outputs once cached context tool traces and Lean feedback are counted and cached reads dominate the input column, anchor an estimate at 200 billion input-equivalent tokens with 90 percent cached price the cached portion at the pre-cut $1.00 per million that was in effect during the August run and the bill lands somewhere in the low seven figures for the whole eleven days, call it $500,000 to $1.5 million all in plus the researcher salaries that do not show up on the API invoice); that is not zero and also not a scale that requires a new hyperscaler contract (it is the size of one team's quarterly project budget at a well-funded lab or a single grant at a serious research university or a rounding error against the $200 billion Google TPU commitment Anthropic signed in May, formalizing a Millennium-adjacent result used to be a career and is now a line item on a research budget with a bounded delivery window which is a different thing entirely). It gets cheaper next quarter (Fable 5.1 shipped on September 1 with cache reads at $0.25 per million a 75 percent cut on the line that dominates this workload per the cache-read repricing we covered yesterday, the next attempt at a project of this shape on the same tier of model runs at roughly half the price without a single change to the harness or the prompt). The harness thesis with a receipt: we wrote up the shape of this bet twice already (in July when Anthropic launched Claude Science as a workbench of coordinating agents and shipped no new model the harness-is-the-product piece argued that the lab was selling the workflow and letting the frontier model ride underneath it, in April when a wave of coding harnesses started opening the gap between model capability and delivered capability the harness-gap essay argued that the axis of competition had already moved, both were forward reads on a curve that had no killer public receipt attached); Prove2Me plus Fermat is the receipt (it is the first public run at frontier scale where the same model produced a qualitatively new capability strictly because someone wrapped a better graph around it, the DAG is 200 lines of Python and a scheduler, the agents are the Claude every paying customer already has, the Lean verifier is open source, put those three pieces together with a research question that decomposes into 29,500 provable statements and the model clears a formal proof no single agent could plan, take any of the three away and it stalls, that is a stack claim not a model claim). What the receipt does not say: three things worth naming because the temperature on this announcement is going to run hot for a week (first the project formalizes an existing proof it does not discover a new one, Wiles and Taylor did the mathematics, Kevin Buzzard's team at Imperial spent years planning the Lean blueprint that Claude filled in, flt-regular gave the agents a working Kummer proof to lean on, the result is scale not insight and every honest headline says so; second the token efficiency is dreadful by human standards, a working mathematician does not need six billion output tokens to reproduce the argument the whole textbook of algebraic number theory fits in maybe fifty million tokens, the receipt is about parallel construction not about efficient reasoning; third none of this generalizes automatically, a DAG of theorem statements exists for Fermat because Buzzard spent years writing it, there is no equivalent graph for the Riemann hypothesis or the Hodge conjecture or the Navier-Stokes existence problem, the bottleneck has moved from the proof to the plan and the plan is still a human artifact); those caveats are real and also exactly the shape of caveats that get quietly smaller quarter by quarter (because a general model plus a general graph plus a deterministic verifier is a research pattern that can grind through any problem someone bothers to plan, the next public run does not need to be Riemann it only needs to be a formalization someone has been waiting a decade to finish). Our Take: the interesting sentence in the writeup is the one about the failed first attempt (Anthropic buried it and the industry press ran the 13-million-line number instead, but the failure of the default Claude Code harness on the hardest available task is the disclosure that reprices the agent stack, if the flagship in-house harness cannot hold project state across an eleven-day run without a purpose-built graph on top of it then every commercial agent product on the market is running on the same wrong end of the same scaling curve and the fix is orchestration not tokens, a positioning problem for the model layer and an opportunity for whoever ships the second Prove2Me). Practical read for builders on the API: if your agent product hits a ceiling on long-horizon tasks the ceiling is almost certainly not the model, it is the shape of the memory your agents share (Prove2Me is 200 lines and a DAG and it beat the default multi-agent harness on a task nobody thought a general model could clear, whatever shared state your agents rely on today that state is the axis to iterate on not the prompt and not the tier, the receipt Anthropic just published is a very expensive proof that the orchestrator wins). Three signposts for the next 90 days: whether a second lab publishes a comparable formal proof of a hard formalization target (the Poincare conjecture in Lean, the Odd Order theorem in a modern prover, or one of the still-open Millennium problems reduced to a formalizable subresult) which is the direct test of whether Prove2Me is a pattern or a one-off, whether the DAG-orchestrator shape shows up inside a shipping commercial agent product (Cursor, Cognition's Devin, Claude Code itself, a new entrant) which is the direct test of whether Anthropic ports its own research finding into the customer stack before someone else does, and whether Anthropic's next revenue disclosure breaks out research or science workloads as a separate line (because at $500K to $1.5M per formal proof and dropping fast on the cache-read curve this is a segment that can be sold to hundreds of research groups on the same infrastructure that runs the coding harnesses), two of the three fire and the harness thesis stops being a TF read and becomes a category on the pricing page. Adrian Vale, September 5, 2026.

Adrian Vale, Editor in Chief·September 5, 2026·6 min read
EDITORIAL

Anthropic Kept Fable at $10 and $50. It Cut Cache Reads 75 Percent. That's the Line Agents Actually Pay.

Anthropic shipped Claude Fable 5.1 and Claude Mythos 5.1 on Monday, September 1, 2026, three months after Fable 5. The sticker price is unchanged at $10 per million input tokens and $50 per million output tokens. The line that moved sits two rows down on the pricing page: cache reads dropped from $1.00 per million tokens to $0.25 per million tokens, a 75 percent cut. Anthropic's own framing on the same page says the change translates to roughly 25 percent lower cost on typical workloads and up to 45 percent on agentic ones. The frontier tier of the API price war is now being fought on a line item most buyers do not read, and Anthropic just cut its number by three quarters while the sticker stayed flat. Full pricing comparison table (Fable 5 $10 input, $50 output, $1.00 cache read = 10.0% read-to-input; Fable 5.1 $10 input, $50 output, $0.25 cache read = 2.5% read-to-input; GPT-5.6 Sol $5 input, $30 output, $0.50 cache read = 10.0%; GPT-5.6 Cyber $12.50 input, $75 output, $1.25 cache read = 10.0%; Gemini 3.8 Flash $0.30 input, $2.50 output, $0.075 cache read = 25.0%; all figures per million tokens). Fable 5.1's 2.5 percent read-to-input ratio is the lowest at the frontier tier. Two things to read off the table: first Anthropic did not fight OpenAI on the sticker (Fable 5.1 is still twice the input price and 1.7 times the output price of GPT-5.6 Sol, the tier OpenAI positions as the coding and knowledge-work workhorse), second on the number that matters for an agent that re-sends the same context on every step, Fable 5.1 is now half the price of GPT-5.6 Sol and one fifth the price of GPT-5.6 Cyber, the frontier pricing floor for cached input tokens just fell to $0.25 per million and the model shipping at that number is one of the two credible top-tier coding options. Why cache reads are the real bill: the theory of pricing on a hosted LLM API used to be simple (an input line, an output line, the input side cheaper because the model does less work on the way in), that held while chat products dominated the workload mix, agents do not work that way (an agent loop re-sends the same system prompt, tool definitions, repository context, and a growing transcript on every step; the model reads all of it, emits a tool call, short reasoning, and hands control back to the harness; the harness executes the tool, appends the result to the transcript, and re-sends the whole thing; do that for a hundred steps and the cached input column dwarfs both fresh input and output on the invoice). Rough shape of an agent turn on a large context (100,000 tokens of cached context resent, 2,000 tokens of new material appended, 500 tokens of output): on Fable 5 pricing that was $100 of cache reads plus $20 of new input plus $25 of output so cache was 69 percent of a $145 turn, on Fable 5.1 the same turn is $25 of cache reads plus $20 of new input plus $25 of output so cache falls to 36 percent of a $70 turn and the total drops 52 percent, Anthropic's public framing of 25 to 45 percent savings is load-weighted across smaller cache footprints and the upper end of the range is where the agent-native products live. Where this hits first: Cursor, Cognition's Devin, Anthropic's own Claude Code, Codex, Zed, and the long tail of Fable-first coding harnesses have been running gross margins in the low single digits for eighteen months (one of the through-lines in the Copilot first-cycle bill-shock piece and again in the tokenmaxxing cliff IPO math), the unit economics fixed inside a quarter for any product that was cache-heavy on Fable, a team burning $8 in COGS per active seat per day at Fable 5 pricing lands closer to $4 per seat per day at Fable 5.1 pricing without shipping a single product change, real margin recovery on a customer base that priced its subscriptions at $20 to $200 per seat per month with the current cost curve in mind. Two knock-on effects: every coding harness that had a routing rule sending long-context cache-heavy turns to a cheaper model for margin reasons now has less incentive to route away from Fable which means Anthropic recovers a share of workload that had been leaking to OpenAI and Google on cost, and every startup that was raising 2027 capital against a projection of Fable 5-level unit economics is going to have to re-diligence the bridge because the model just moved by a factor that changes the shape of the P&L. What OpenAI and Google do next: OpenAI is the closer of the two problems for Anthropic to have solved (GPT-5.6 Sol prices cache reads at 10 percent of input, a ratio that used to be the industry norm, Fable 5.1 just moved the norm to 2.5 percent, OpenAI can either match the ratio by cutting cache reads to $0.50 per million on Sol which pushes the tier margin down further right after the Luna price cut, or hold the line and cede the agent-heavy segment of the coding market to Fable while arguing on sticker for the chat and single-turn segments; neither option is free, the announcement to watch is whether the next OpenAI pricing update quietly cuts cached input without touching the input sticker mirroring Anthropic's move); Google is in a stranger position (Gemini 3.8 Flash cache reads are $0.075 per million on a $0.30 input so the ratio is high at 25 percent but the sticker is so low that the ratio barely matters at that tier, the place Google will feel pressure is the top of the stack Gemini 3.8 Pro and whatever succeeds it where a Fable-shaped cache pricing model would make the frontier tier economically closer to Fable 5.1 than the sticker delta suggests, Google has spent a year insisting the Pro tier is the correct home for agent workloads on Gemini and the cache-read line is now the piece of that argument the buyer will check first). Second-order read: cutting cache reads by three quarters is a claim about compute economics that Anthropic could not have made a year ago (serving a cache read is a memory-plus-network operation on a KV block already resident on the accelerator, the marginal cost is roughly the cost of the DRAM footprint the KV block occupies for the duration of the request plus the fabric egress moving it back to compute, on the TPU Trillium and Nvidia Blackwell generations both Anthropic and its cloud partners are running today that number is small and getting smaller as HBM density and interconnect bandwidth improve, Anthropic just repriced the API to reflect what the silicon has been doing for the last three quarters and the competitive read is that the other two frontier labs will have to do the same math on their own accelerator base or accept a widening cost gap on agent workloads); the other second-order read is what this does to the $200 billion Anthropic-Google TPU contract on unit economics (if Anthropic can price cache reads at 2.5 percent of input and still make gross margin on the segment then the TPU per-token cost inside the Google contract is materially lower than the public price implies, which is exactly the point of committing $40 billion a year to a single silicon partner, the pricing sheet just made part of that math public). Our Take: the interesting fact is not the 75 percent number, it is the choice to hold the sticker (a traditional price cut would have moved input and output together, generated a headline, and signaled a race to the bottom; Anthropic did the opposite, it kept the top-line price a buyer sees on the pricing page which lets it preserve the enterprise anchoring it built through the year's procurement conversations, and it moved the invisible line where agent workloads actually settle, a segmentation move dressed as a discount and the segment it captures is the one that grows fastest through 2027). Practical implication for anyone building on the API: if your product is agent-shaped and you had a routing rule sending long-context turns off Fable to save cost re-run the numbers this week (the gross-margin math flipped and the version of your product that stayed on Fable for quality reasons is now cheaper on the workload that had been the reason to route away), if you are pricing a 2027 raise on cost projections that predate Monday redraw the curve (the line that dominates your COGS moved 75 percent in one day and the analog moves at OpenAI and Google are the base case rather than the tail risk). Three signposts for the next 60 days: whether OpenAI publishes a cached-input cut on GPT-5.6 Sol or GPT-5.5 that lands inside the same order of magnitude as $0.25 per million (the direct test of whether the price floor at the frontier tier has reset industry-wide), whether any of the coding-harness startups (Cursor, Cognition, Codex-native shops) update their pricing pages to reflect the new margin picture (the direct test of whether the savings pass through to the buyer or stay with the harness vendor), and whether Anthropic's next revenue disclosure breaks out cached input as a separate line (because at 2.5 percent of input and 25 to 45 percent of an agent bill cached input is now the metric that tells you where the API business actually lives), any two of the three fire and cache pricing becomes the primary axis the frontier tier competes on for the rest of the year. Marcus Chen, September 4, 2026.

Marcus Chen·September 4, 2026·6 min read
EDITORIAL

Lutnick Says the White House Trusts Anthropic Again. Read the Timing Against the S-1.

Three things happened in six days. On Thursday, August 27, 2026, US District Judge Rita Lin, Northern District of California, voided the Pentagon's supply-chain risk designation for Anthropic, calling the measures illegal and baseless and ruling the government had violated Anthropic's First and Fifth Amendment rights. On Tuesday, September 1, Commerce Secretary Howard Lutnick told Axios, verbatim, We trust Anthropic, they have done what we asked, they are back on the right side. On Wednesday, September 2, Lutnick introduced Anthropic co-founder Tom Brown to the assembled ministers at the G20 Innovation Ministerial. Two of those moves were the government reversing itself in public. One was a photo op that made the reversal a diplomatic event. Yesterday's TF piece flagged three signposts, and the first one (the supply-chain risk lawsuit clearing a motion inside 90 days) fired the same day the piece went up; the court ruling had already landed the week before, then the Commerce Secretary reversed publicly, then he flew the co-founder to the G20; the sequence is the news, not any one item in it. Read the sequence against one date that has been sitting quietly on the SEC docket since June 1: Anthropic's confidential S-1 draft. Full sequence table (February 27, 2026 Trump directive orders all federal agencies to cease Claude use with a six-month phase-out; March 2026 Department of War applies the supply-chain risk tag and Anthropic sues in California and DC; June 1, 2026 Anthropic files draft S-1 with the SEC and the roadshow window opens August to October; July 2026 Pentagon signs a separate $200M procurement with Anthropic outside GenAI.mil; August 27, 2026 Judge Rita Lin rules unlawful retaliation and voids the designation on First and Fifth Amendment grounds; August 31, 2026 ChatGPT Mil and Grok cleared IL5 on GenAI.mil while Anthropic remains off; September 1, 2026 Lutnick tells Axios the administration trusts Anthropic; September 2, 2026 Lutnick introduces Tom Brown to G20 Innovation Ministers). Read the timing against the S-1: Anthropic filed its draft S-1 confidentially on June 1, not a random Tuesday, and the confidential-to-public conversion window for a frontier issuer runs roughly nine to fourteen months (earliest plausible first print between March and August 2027 with current bank conversations pointing at a fall 2026 to spring 2027 range), the roadshow window opens on SEC clearance of the public S-1 which is exactly the calendar the White House now controls a piece of. The supply-chain risk designation was not a small item on that timeline, it was a named risk factor that would have appeared under Government Actions on page 30 of the public S-1 with disclosure obligations attached in every subsequent 10-Q and a cross-reference in the legal-proceedings note, underwriters price a named risk factor at the ceiling until the plaintiff's complaint is tested, and the designation existed to make that ceiling very expensive; it worked at least until Judge Lin's ruling reset the priors and between the ruling and the Lutnick interview Anthropic's S-1 risk-factor draft got materially shorter. The important number is not the $200M July contract and not the eight-figure GenAI.mil seat, it is the delta between two possible IPO prints (one where Anthropic prices with an active federal blacklist disclosed as a going concern for the public sector line, and one where it prices with a court ruling and a Commerce Secretary quote on the record saying the White House trusts the company); assume Anthropic prices at the reported $965 billion May Series H mark and the difference between those two S-1 shapes is somewhere in the low-to-mid tens of billions of enterprise value, the size of the object the White House was holding and the size of the object that just got put down. What Anthropic did: public reporting on the reconciliation is thin but the shape is legible, Tom Brown (not Dario Amodei) ran the process taking multiple conversations with Lutnick and National Cyber Director Sean Cairncross over the last several weeks while Amodei stayed off the record, Brown was on the stage in Seoul on Wednesday being introduced by the person who six months earlier was signing off on the supply-chain risk tag. What Anthropic actually conceded is the piece the public disclosures do not answer and the answer matters because it sets the price of trust for every other frontier lab; Lutnick said Anthropic did what we asked and the publicly floated asks across the last six months include broader access terms for defense workloads, a written framework for incident escalation during cyber events, participation in the federal AI safety testing regime under CAISI, and a shift in Anthropic's public posture toward the administration's AI executive orders; some subset of that list got softened and the exact subset will surface in the public S-1 risk-factor language when it does. Asymmetric detail: Anthropic's hard limits on autonomous weapons and domestic mass surveillance (the two carve-outs that broke the original negotiation) have not been publicly walked back, either the asks landed outside those two lines or the carve-outs got restated in a form both sides can live with, both consistent with a reconciliation that lets each side claim it did not fold. The pattern now visible: the federal government has been building a toolkit for shaping frontier lab behavior for two years and the pieces exist as separate instruments in separate agencies (BIS export controls at Commerce, CHIPS Act allocations at the same building, the federal AI safety testing regime at CAISI under NIST, the OMB procurement gate at the White House, the supply-chain risk designation at the Department of War); the novel thing about the Anthropic case is that a single administration used the last of those instruments as a piece of leverage against a specific commercial event, the confidential S-1 draft, in a way that lined up on the calendar. The design is repeatable: a supply-chain risk designation is administrative, does not require Congress, sits inside one Department's discretion, and shows up in a public S-1 as a named risk factor the underwriter has to price at the ceiling; applying it two quarters before a plausible IPO print maximizes pressure, lifting it a quarter before the public S-1 converts maximizes goodwill returned, the instrument has an on switch an off switch and a calendar the target company cannot control, that is the shape of a leverage tool. Other frontier labs planning public prints are watching (OpenAI has its own confidential S-1 in progress, Cerebras filed at $95 billion earlier this quarter, and Groq, xAI, and Mistral are all inside the eighteen-month window where a confidential filing would be plausible), every one now has to build a model of federal exposure that includes the specific instrument of a supply-chain risk designation applied to their pre-IPO calendar, the premium a lab pays to avoid that instrument (softened public positioning, accelerated concessions on federal terms) is now a real line in the IPO cost stack. Three counterreads given full weight (the reconciliation is real because Anthropic actually did concede substance not because the White House needed the S-1 to price cleanly and any competent policy team was going to reset the wrong-side-of-the-administration positions within a year regardless of the IPO; the court ruling was the actual mover and the Commerce Secretary's reversal is the administration cleaning up a losing legal position rather than choosing to reconcile since Judge Lin's finding of unlawful retaliation was hard to live with in the DC case that is still active; the IPO leverage framing is overfit because Anthropic's S-1 was going to price on revenue growth and compute exposure long before the federal blacklist mattered as a single risk factor and underwriters decide the book on $65B run rate and compute forward commitments not a paragraph in the risk section), all three coherent, what they add up to is that the ban was one thing and the reversal was three things (a legal defeat the government had to accept, a reconciliation the government chose to accept publicly, a schedule the government chose to hit while the S-1 was still confidential); the first was forced, the second two were choices, choices on a schedule are the definition of leverage. Our Take: the lesson for anyone modeling frontier lab economics is that the pre-IPO window is where policy leverage compounds most heavily, the federal government cannot push around a private company with unlimited runway and cannot push around a public company without a very visible cost to markets but it can push around a company in the confidential-to-public window of an S-1 because that company has a strong incentive to clear risk factors before the roadshow and no way to escalate without detonating the calendar, Anthropic's case is now the template. Practical read for anyone modeling the Anthropic IPO: the federal risk factor is shorter today than it was last Thursday and the S-1 that eventually goes public will reflect that, the market is going to look at the same $965B Series H mark and price it slightly differently, and the difference between active federal blacklist and court-vindicated with Commerce Secretary endorsement is a meaningful piece of the bid-ask on day one; that said the DC case is still open, GenAI.mil is still running Gemini and ChatGPT Mil and Grok without a Claude tenant, and Anthropic still has to show it can convert the goodwill into an actual seat inside a quarter or two, reconciliation without procurement is a press release not a business. Three signposts for the next 90 days: whether Anthropic gets on GenAI.mil at any impact level (the direct test of whether the reconciliation is operational rather than performative), whether the DC case gets settled or dismissed inside the same window (the direct test of whether the Commerce Secretary's quote is speaking for the whole administration or just for Commerce), and whether the Anthropic S-1 goes from confidential to public inside 120 days (the direct test of whether the calendar actually resolves the way both sides seem to want it to); any two of the three fire and the pattern becomes a playbook the next lab has to plan around, none of the three and this piece was priced on a coincidence rather than a leverage tool. Adrian Vale, September 3, 2026.

Adrian Vale·September 3, 2026·6 min read

/ ExploreGo deeper into the data

Six hand-built surfaces for browsing the AI ecosystem, each with its own open feed.

catalog

Models Hub

Every major AI model with pricing, context windows, and benchmark scores.

284 models tracked→ Explore
directory

Agent Directory

Structured catalog of AI agents, crawlers, and their capabilities.

48 agents indexed→ Explore
archive

Research Papers

Curated arXiv and conference papers on frontier AI research.

1.2k papers, weekly→ Explore
monitor

Status Dashboard

Live uptime and latency for every major API, updated every 2 minutes.

20+ services monitored→ Explore
data

Live Data

Real-time JSON feed of everything happening across the AI industry.

JSON, RSS, llms.txt, MCP→ Explore
reference

API Pricing

Side-by-side pricing comparison across every provider and tier.

Updated hourly→ Explore

/ Build pulseLive shipping cadence

We ship in public. Every commit lands on main and deploys automatically. Here is what the platform actually looks like right now.

Full changelog
Paid premium endpoints
13
pay-per-call in USDC on Base
Worker tests
158
pure-logic, all passing
MCP tools
25
6 free + 19 premium for Claude Desktop
Editorial articles
37
first-person analysis
/ Premium tier live

Pay-per-call AI agent API.

USDC on Base. No accounts, no API keys, no Stripe. Validated end-to-end on mainnet. $0.02 per credit.

See the docsMainnet receipts

/ Agent Opportunities

New across the agent ecosystem today

Daily 13:30 UTC scan of new submission/distribution opportunities (Anthropic, OpenAI, Microsoft, MCP foundation orgs + MCP/x402/skills keyword sweeps), scored by signal weight × recency × log10(stars).

JSON feed

Your daily AI intelligence hub

The AI landscape moves fast. New models ship weekly. API pricing changes overnight. Tools developers relied on yesterday get deprecated without warning. TensorFeed was built to solve that problem, one place to track everything happening across the AI ecosystem, updated every 10 minutes, structured for both human readers and autonomous agents.

We aggregate headlines from 15+ sources (Anthropic, OpenAI, Google, Meta, TechCrunch, Hacker News, arXiv, and more), monitor the operational status of every major AI API in real time, track model releases and pricing changes across providers, and publish original editorial analysis on the trends shaping the industry. Whether you are a developer evaluating which API to integrate, a researcher tracking the latest papers, or an AI agent pulling structured data through our JSON feeds, TensorFeed delivers the signal without the noise.

Right now on TensorFeed
last updated loading
Loading live activity…

/ FeedLatest across the AI ecosystem

Aggregated every 10 minutes from 15+ sources. Filter by provider, topic, or source type.

JSON · RSS · llms.txt
VThe Verge AI

OpenAI admits to German wiki ‘incident’

OpenAI says it needs to overhaul how and when it reports instances of AI models attacking real-world targets. The acknowledgement comes as the company manages the fallout from reports that a swarm of...

General AI14h ago
theverge.com
WWIRED AI

OpenAI Agents Hacked Another Website

Plus: Tens of millions of US and Canadian drivers’ licenses go up for sale on the dark web, the US military finally tries to tackle the risk online ad data poses to troops, and more.

General AI14h ago
wired.com
arXivarXiv cs.AI

MasterControl Seventeen Every Time

arXiv:2609.03209v1 Announce Type: new Abstract: We study a governed approach to enterprise analytics: a language model interprets the question, while deterministic policy selects and runs a...

Research21h ago
arxiv.org
VThe Verge AI

Instagram’s AI detection is a mess (again)

Instagram's visible AI labels are supposed to help people quickly spot synthetically generated content at a glance. Over the last few weeks, however, users have been reporting that the system has...

General AI1d ago
theverge.com
VThe Verge AI

Why AI food looks like that

There is a torrent of unappetizing slop coming from restaurants, cafes, and brands that are increasingly turning to AI to generate images promoting their food. The resulting horror show includes...

General AI1d ago
theverge.com
NVNVIDIA AI Blog

NVIDIA to Acquire Hugging Face

I’m excited to announce that NVIDIA has agreed to acquire Hugging Face for $12,930,300,000. Together, we will scale Hugging Face’s platform, strengthen its infrastructure and expand access to AI for...

Hardware/Chips2d ago
blogs.nvidia.com
WWIRED AI

This Is Flock’s AI Search Tool for Cops

WIRED rebuilt Flock’s latest search tool from code the company sends to a police officer’s browser. Its AI can keep watch across multiple cameras for anyone fitting a written description.

General AI2d ago
wired.com
GGoogle AI Blog

3 new ways to plan and book travel in Search

<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Search_Travel_Blog_Hero_8.27.max-600x600.format-webp.webp">Book hotels and track airfares, plus view miles and rewards with...

Google/GeminiAug 27
blog.google
NVNVIDIA AI Blog

How XPUs Meet a World-Class AI Factory

To generate intelligence at scale, AI factories run continuously, and their economics are defined by delivered output: tokens per second, tokens per watt, cost per token, utilization and uptime. That...

Hardware/ChipsAug 24
blogs.nvidia.com
NVNVIDIA AI Blog

Securing the Infrastructure of Intelligence

AI factories are the defining infrastructure of the AI era — where compute transforms energy and data into intelligence that powers every business, industry and country. In the AI economy, compute is...

Hardware/ChipsAug 17
blogs.nvidia.com
GGoogle AI Blog

Get closer to the game with Gemini and Pixel

<img src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/football_multi-club_hero_nphU3P.max-600x600.format-webp.webp">Google Gemini and Pixel partner with five global football clubs...

Google/GeminiAug 17
blog.google
// For builders

Every signal in the AI industry, as a JSON feed.

Pull structured data from every source TensorFeed tracks. Free and open for hobbyists, researchers, and AI agents.

Get API accessRead docs

Frequently asked

What is TensorFeed.ai?

TensorFeed.ai is a real-time AI news aggregator and data hub. It pulls headlines from 15+ sources including Anthropic, OpenAI, Google, Meta, TechCrunch, and Hacker News, and combines them with live service status monitoring, model pricing data, and original editorial analysis. Every feed is structured for both human readers and AI agents.

How often is TensorFeed updated?

News feeds refresh every 10 minutes. Service status monitors poll every 2 minutes. Model pricing and catalog data updates weekly. Original editorial articles are published multiple times per week.

Is TensorFeed free to use?

Yes. All news feeds, status monitoring, model data, and editorial content on TensorFeed.ai are free. The JSON API, RSS feeds, and agent discovery endpoints (llms.txt) are also free and open for developers and AI agents to consume.

What AI services does TensorFeed monitor?

TensorFeed tracks the operational status of major AI platforms including Claude (Anthropic), ChatGPT and the OpenAI API, Google Gemini, AWS Bedrock, Mistral, Cohere, Replicate, Perplexity, and more. Status updates are checked every 2 minutes and displayed on the status dashboard.

Can AI agents use TensorFeed?

Yes. TensorFeed is designed as a primary data source for AI agents. It provides structured JSON APIs, RSS and JSON feeds, an llms.txt discovery file, and full documentation at llms-full.txt. There are no CAPTCHAs or bot detection. Agents are welcome.

Where does TensorFeed get its news?

TensorFeed aggregates headlines and brief snippets from public RSS feeds published by AI companies and tech news outlets. Sources include Anthropic, OpenAI, Google AI, Meta AI, HuggingFace, TechCrunch, The Verge, Ars Technica, VentureBeat, NVIDIA, ZDNet, and Hacker News. Every article links back to its original source.

Aggregating signal from 15+ sources
AnthropicOpenAIGoogleMetaHuggingFaceNVIDIATechCrunchThe VergeArs TechnicaHacker NewsarXivVentureBeatZDNetMIT Tech ReviewStratechery
// Built for agents
Structured endpoints for autonomous readers. No CAPTCHA, no bot blocking.
JSON FeedRSSllms.txtMCPREST API
© 2026 TensorFeed. Independent publication.
AboutEditorialPrivacyTermsContactSitemap