FactCheckRadar Real-Time Fact Checking 358 published fact-checks

Built for rumors, fast-moving news, and claims that spread faster than the proof.

Latest fact-checks

Supported

OpenAI Chief Scientist Calls Past Two Years 'Surprisingly Slow' as New Release Pace Sparks Speculation

Claim checked

“Jakub Pachocki “We see pretty significant improvements in the short term, extremely significant improvements in the medium term,” said Jakub Pachocki, OpenAI’s chief scientist. “In fact, I would say, like, I think the last two years have been surprisingly slow.” OpenAI staff said the fast paces of releases should keep coming for the foreseeable future. Meaning we can expect a new model within <60 days”

OpenAI Chief Scientist Jakub Pachocki has characterized the last two years of AI development as "surprisingly slow," promising "extremely significant improvements" in the near future. These comments coincided with the April 2026 launch of GPT-5.5, which arrived just one month after its predecessor, GPT-5.4. OpenAI leadership has officially confirmed that they intend to...

Read full report ↗
Supported

Local AI Beast Awakens: Qwen 3.6 27B + Pi Conquers MacBook Pro Hardware

Claim checked

“Qwen 3.6 27B + Pi on a MacBook Pro, fully local: a beast. 27B dense model, flagship-level agentic coding, running entirely on hardware in your hands. Speed is impressive, Utility is extremely high. It punches orders of magnitude above its weight. Local AI is getting real.”

Claims in the X post about Qwen 3.6 27B running fully locally on a MacBook Pro with Pi, delivering flagship-level agentic coding as a 27B dense model with impressive speed and utility, hold up against recent evidence from official releases and tech reports.

Read full report ↗
Supported

Performance of MiMo V2.5 Pro on agentic and real-world work benchmarks

Claim checked

“MiMo V2.5 Pro leads its peer group on agentic tasks. The model scores 1578 on GDPval-AA, and places it in the top tier for real-world work tasks among recent releases”

The claim that MiMo V2.5 Pro leads its peer group in agentic tasks and ranks in the top tier for real-world work is supported by independent benchmarking data. As of April 2026, the model holds an ELO of 1580 on the GDPval-AA leaderboard, placing it among the highest-performing models globally and leading major competitors in its class.

Read full report ↗
Supported

Claim that DeepSeek-V4-Pro pricing will fall after Huawei Ascend 950 deployment

Claim checked

“"DeepSeek said Pro pricing could fall sharply once Huawei Ascend 950 supernodes are ​deployed at scale in the second half of the year"”

DeepSeek has officially indicated that the pricing for its DeepSeek-V4-Pro model is expected to decrease significantly in the second half of 2026. This reduction is tied to the planned large-scale deployment of Huawei Ascend 950 supernodes, which will address current computing power limitations.

Read full report ↗
Unverified

Fact Check: Were GPT 5.5 and GPT 5.5 Pro spotted on OpenRouter?

Claim checked

“GPT 5.5 and GPT 5.5 Pro spotted on OpenRouter Its coming”

Social media reports claim that OpenAI's unreleased GPT 5.5 and GPT 5.5 Pro models have been spotted on the LLM aggregator platform OpenRouter. While specific model IDs have appeared in the platform's URL structure, the models are currently marked as "not available," and they do not appear in the official public directory.

Read full report ↗
Supported

Fact Check: Pricing of GPT-5.5 compared to GPT-5.4 and Claude Opus 4.7

Claim checked

“GPT-5.5 +100% more expensive than GPT5.4 and +20% more expensive than Claude Opus 4.7! When exactly will token prices hit CPI?”

The claims regarding the pricing of OpenAI's GPT-5.5 are accurate based on API rates released during its launch on April 23, 2026. GPT-5.5's per-token costs are exactly double (+100%) those of its predecessor, GPT-5.4. Additionally, while its input costs match Anthropic's Claude Opus 4.7, its output costs are 20% higher.

Read full report ↗
Supported

Fact Check: Does GPT-5.5 underperform Claude Opus 4.7 on the SWE-Bench Pro coding benchmark?

Claim checked

“GPT 5.5 underperforms Opus 4.7 on SWE-Bench Pro. Couldn't find any reported SWE-Bench scores at all and an internal benchmark is reported instead. That footnote is trying really hard to bury the lede. GPT 5.5 isn't SOTA for coding.”

Evidence from official release notes and independent technical analysis confirms that Claude Opus 4.7 outperforms GPT-5.5 on the SWE-Bench Pro benchmark, a key metric for real-world software engineering. While OpenAI's announcement highlights a state-of-the-art (SOTA) score on a different benchmark (`Terminal-Bench 2.0`), its reported score for SWE-Bench Pro is lower than the...

Read full report ↗
Supported

SpaceX Agrees to $60 Billion Deal with AI Startup Cursor

Claim checked

“SPACEX SAID TO AGREE TO BUY CURSOR FOR MORE THAN $50 BILLION- NYT”

SpaceX has officially announced a partnership with the AI coding startup Cursor that includes a structured acquisition option. According to official statements and reporting from CNBC, The Guardian, and Business Insider, SpaceX has secured the right to acquire Cursor for $60 billion later in 2026. Alternatively, the companies may remain partners with SpaceX paying $10 billion...

Read full report ↗
Unverified

Italy Fast-Tracks Tesla FSD Approval Claims Unverified

Claim checked

“BREAKING: ITALY MOVES TO ACCELERATE TESLA FSD SUPERVISED APPROVAL $TSLA • The Italian Ministry of Infrastructure and Transport has officially forwarded a request to fast-track the approval of Tesla’s Full Self-Driving (Supervised) system. • The request was sent directly to the Transport Department and the relevant Italian DMV-equivalent authorities for review.”

Social media reports claiming the Italian Ministry of Infrastructure and Transport has officially moved to fast-track Tesla's Full Self-Driving (Supervised) system are currently unverified. While the Netherlands recently became the first European country to approve the system, there is no official confirmation from Italian authorities regarding a specific "fast-track" request.

Read full report ↗
Supported

Moonshot AI Launches Kimi K2.6 with Top Coding and Agent Benchmarks

Claim checked

“Kimi K2.6 Launched: Open-Source + Competing with Frontier Models - Agentic coding king: #1 on SWE-Bench Pro (58.6), beating GPT-5.4 xhigh (57.7), Gemini 3.1 Pro (54.2), Claude Opus 4.6 (53.4)”

Moonshot AI has officially released Kimi K2.6, an open-weight model designed for complex agentic tasks and long-horizon coding. Evidence confirms it achieved a score of 58.6 on SWE-Bench Pro, placing it at the top of the leaderboard and surpassing models like GPT-5.4 and Claude Opus 4.6 in that specific metric.

Read full report ↗
Supported

Oxford Professor Michael Wooldridge Claims ChatGPT Lacks True Understanding

Claim checked

“Oxford AI professor Michael Wooldridge: "ChatGPT doesn't understand anything. It's essentially doing some fancy statistics."”

Michael Wooldridge, a professor of computer science at the University of Oxford, has consistently argued that Large Language Models (LLMs) like ChatGPT do not possess genuine understanding. He describes their functionality as sophisticated statistical pattern-matching rather than a grasp of meaning or real-world context.

Read full report ↗
Supported

OpenAI 'Spud' Model Confirmed for Release in April 2026

Claim checked

“'Spud ' confirmed for this week. As it is written, so it shall be done. BUCKLE. UP.”

Multiple reports and statements from OpenAI leadership confirm that a new frontier model codenamed 'Spud' has completed pre-training and is scheduled for release in the second quarter of 2026. OpenAI President Greg Brockman recently confirmed the model's existence, describing it as a new foundational base model resulting from two years of research.

Read full report ↗
Mostly true

GPT-5.5 'Spud' in Testing and Comparison to Claude Opus 4.7

Claim checked

“Yep. GPT-5.5 has been in testing on ChatGPT for a few days now, and I can confirm it’s better than Opus 4.7.”

While evidence confirms that GPT-5.5 (code-named "Spud") is a real model nearing release and Claude Opus 4.7 was recently launched, the claim that GPT-5.5 is currently being "stealth tested" on ChatGPT is unverified. GPT-5.5 Status: OpenAI executives confirmed pretraining was completed in late March 2026, with a release expected in Q2 2026. Opus 4.7 Status: Anthropic released...

Read full report ↗
Supported

FBI Raids Home of Suspect in Sam Altman Molotov Cocktail Attack

Claim checked

“FBI agents raided the home of the 21-year-old suspect accused of throwing a Molotov cocktail at Sam Altman’s home. He was reportedly driven by anti-AI views and carried a manifesto listing AI executives and investors with addresses.”

On April 13, 2026, FBI agents raided the Texas home of Daniel Moreno-Gama, a suspect accused of throwing a Molotov cocktail at the San Francisco residence of OpenAI CEO Sam Altman. Investigators confirmed the suspect was motivated by anti-AI views and carried a manifesto that included a "hit list" of other AI executives and investors along with their home addresses.

Read full report ↗
Mostly true

Kimi K2 Model Released by Moonshot AI

Claim checked

“Kimi K2.6 code preview is ou”

Moonshot AI has released Kimi K2, a large-scale Mixture-of-Experts (MoE) language model with 1 trillion total parameters. While the user claim mentions a "Kimi K2.6 code preview," official documentation and repository data currently identify the latest major release as Kimi K2, which includes specialized variants for base research and instruction-based agentic tasks.

Read full report ↗
Mostly true

DeepSeek V4 Release Timing and Huawei Delay Claims

Claim checked

“From an interview with a mysterious source: DeepSeek V4 is indeed scheduled to be released next week. The delayed release of DeepSeek V4 has nothing to do with Huawei Ascend; it is purely because the results were not satisfactory.”

While reports from mid-April 2026 support a potential release of DeepSeek V4 in late April, the claim that delays are unrelated to Huawei Ascend hardware is contradicted by multiple industry reports. Release Timing: Sources indicate a target window of late April 2026, which aligns with the "next week" claim made on April 19. Huawei Involvement: Major reporting from the...

Read full report ↗
Mostly true

DeepSeek-V4 Release Rumors and Technical Specifications

Claim checked

“Okay, Deepseek V4 is coming out this week! 1.6T MOE model, a competing model to Opus-4.6/GPT-5.4 MMLU 99.4% / SWE 83.7% Since the parameters are so large, only the quantized version will be possible on a 512GB Mac. Sorry it's a Korean article, you'll have to translate it to read”

Reports and social media posts from April 20, 2026, suggest that DeepSeek-V4 is slated for release as early as this week. While technical specifications such as a 1.6 trillion parameter MoE architecture and high benchmark scores (MMLU 99.4%) are circulating, these figures originate from unofficial leaks and researcher predictions rather than a formal announcement from...

Read full report ↗
Unverified

OpenAI Employee's Description of Codex as 'Proto-AGI' is Unverified

Claim checked

“Codex, and what Codex becomes in the near term, to me is a proto-AGI. We’re on the flight path, we’ve been descending down through the clouds for a while and we can now see visible landmarks — some familiar and some we’ve never seen (or never seen from this POV). What a time to be alive. (This is not a formal OpenAI POV, rather just the musings of some random head of broccoli)”

The claim that an OpenAI employee described the Codex model as a "proto-AGI" is unverified based on the available evidence. While records confirm that the author, Adam.GPT (@TheRealAdamG), is a Go-To-Market (GTM) employee at OpenAI and was posting about Codex on April 18, 2026, the specific text of the statement is not contained within the parsed content of the provided...

Read full report ↗