Google DeepMind ships Gemini 4 Argon; Perplexity adds it for Pro and Max
10/16/2026 — 10/16, 23:20·2 sources (all)·2 reports (all)
Key takeaways
- Google DeepMind's Gemini 4 Argon scored 97.5% on ARC-AGI-3, making it the second lab past 95% after OpenAI, though its score falls to 49.8%…
Background & analysis
On October 16, 2026, Perplexity announced on its official X account that Google DeepMind's Gemini 4 Argon is now available in Perplexity, selectable by Pro and Max subscribers. The post said the model sets a new record on WANDR, but gave no specific score, testing conditions, or other availability details.
The same day, The Decoder, citing Epoch AI, reported that Gemini 4 Argon scored 97.5% on ARC-AGI-3, making Google the second lab after OpenAI to cross the 95% mark, behind only GPT-6 Astra at 99.9%. The report also noted that under the standard harness the score drops to 49.8%, and that Argon still trails rivals in math and software engineering. It added that reaching that ARC-AGI-3 result cost $4,800.
So far, Perplexity has only stated that the model is live and that it set a WANDR record, without disclosing score details. The Decoder and Epoch AI, meanwhile, present two markedly different ARC-AGI-3 results. The situation rests there, with no further official explanation yet.
AI-generated from 2 reports · updated 49 minutes ago
Latest turnAccording to The Decoder, citing Epoch AI, Google DeepMind's new frontier model Gemini 4 Argon scored 97.5% on ARC-AGI-3, making Google the second lab after OpenAI to pass the 95% mark and trailing only GPT-6 Astra at 99.9%. The score drops to 49.8% under a standard test harness, and Argon still lags rivals in math and software engineering.
Related AI tools
- Heat index
- 194
- All sources
- 2
- All reports
- 2
- First seen
- 1 hour ago
Reports on this story headlines open the original
According to The Decoder, citing Epoch AI, Google DeepMind's new frontier model Gemini 4 Argon scored 97.5% on ARC-AGI-3, making Google the second lab after OpenAI to pass the 95% mark and trailing only GPT-6 Astra at 99.9%. The score drops to 49.8% under a standard test harness, and Argon still lags rivals in math and software engineering.
The DecoderAI score 65Perplexity announced on its official X account that Gemini 4 Argon is now available to Pro and Max subscribers, saying the model sets a new record on the WANDR benchmark. The post did not include a specific score, testing conditions or other availability details.
X：Perplexity (@perplexity_ai)AI score 56
Other stories people are talking about
- Heat index 75Tutorial: isolating AI referral traffic in GA4 and BigQuery1 sources
- Heat index 267Gemini 4 Argon Rolls Out to AI Pro Users in Antigravity5 sources
- Heat index 107Perplexity open-sources SPACE agent sandbox and publishes escape tests2 sources
- Heat index 67Hands-on comparison of Gemini 3.1 Pro, Claude Opus 4.6 and GPT-5.3 Codex1 sources
- Heat index 79GEO tips: getting AI assistants to recommend your local business1 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion