Skip to content
Model Season every model has a seasonRadar
Radar · Oct 5, 2026

OpenAI safety employee resigns, saying the company's culture is broken

7 itemsRSS

News × data

The story published out there and, below it, what real token traffic from the last 7 days says about who is in it. The headline opens the original article; the model name opens its page here.

Safety ·

OpenAI safety employee resigns, saying the company's culture is broken

According to TechCrunch, David Robinson left OpenAI and said the company's "culture is broken." Robinson admits he is "something of a cliche": an employee at a leading AI company who issues a dire warning while resigning. The Guardian and The Atlantic carry the same assessment.

Market ·

OpenAI launches visual ads in ChatGPT, initially alongside image generation

According to TechCrunch, the new visual ads appear after image generation results, in the U.S. only for now. The Verge reports the format puts images of sponsored products and services on screen, with testing in the US starting later this month. OpenAI is also expanding measurement tools for advertisers.

Regulation ·

OpenAI adds text watermarking to ChatGPT and Codex, starting in the European Union

According to The Verge, an invisible, machine-readable watermark is rolling out to ChatGPT and Codex text output, at first only for users in the European Union. OpenAI says its textGrain approach "matched or exceeded" others such as Google DeepMind's SynthID. OpenAI says access to detection starts with researchers.

Changes in the data

Price · DeepSeek

DeepSeek V4 Flash 0731 got 246% more expensive

From $0.096 to $0.331 per 1M tokens, blended. Input $0.021 → $0.015, output $0.32 → $1.28. New price since Oct 4.

What the data says. The model has 3.61% of traffic over the last 7 days, ranked 7th. DeepSeek has 23.5% of tokens and 12.6% of estimated spend this week. The lab's most used model is DeepSeek V4.1 Flash, with 15.44% at $0.53 per 1M, and this one costs 1.6 times less.

See DeepSeek V4 Flash 0731 on Model Season →
Removed · OpenAI

Sora 2 Pro was removed from the catalog

Last listed on Oct 4, 2026. Anyone calling this identifier needs to migrate.

What the data says. OpenAI has 10.6% of tokens and 31.7% of estimated spend this week. The lab's most used model is GPT-6 Luna, with 3.73% at $0.20 per 1M.

Price · Meta

Llama 3.3 70B Instruct got 87% more expensive

From $0.15 to $0.29 per 1M tokens, blended. Input $0.10 → $0.22, output $0.32 → $0.50. New price since Oct 4.

What the data says. No volume over the last 7 days. Its weekly peak was 6.64%.

See Llama 3.3 70B Instruct on Model Season →
Price · Qwen

Qwen3.8 27B got 10% cheaper

From $1.06 to $0.96 per 1M tokens, blended. Input $0.42 → $0.42, output $3.00 → $2.55. New price since Oct 4.

What the data says. The model has 0.32% of traffic over the last 7 days, ranked 37th. Qwen has 0.9% of tokens and 0.5% of estimated spend this week.

See Qwen3.8 27B on Model Season →

In the news: summaries written by AI using only what the cited sources published, reviewed before going live. The numbers in the “The data” strip come from our data, traffic from the last 7 days, never from the summary. Catalog snapshot of Oct 5, 2026, compared with the previous ones since Sep 28, 2026. Traffic through Oct 4, 2026. Price is a 75% input and 25% output blend, and only becomes news when the new value holds for two snapshots. Order follows a fixed relevance formula that weighs the type of change and the model’s share.