News archive
All posts and signals, newest first.
SearchMeta released Muse Spark 1.3
• beats GPT 5.6 Sol and Opus 5 on long coding tasks • matches GPT 5.6 Sol on terminal coding • huge jump in long context, scoring 98%+ • uses around…
Google released Gemini 3.8 Flash
• still $0.75 per 1M input tokens • at the same price, it is noticeably stronger than Gemini 3.7 Flash across almost the entire table • especially…
Anthropic released Claude Fable 5.1
The benchmarks show a pretty solid upgrade, with the model beating the previous version and most of its main competitors across the majority of…
CLAUDE FABLE 5.1 IS OUT
Claude Fable 5.1 just showed up in the docs
Nothing special, GPT Astra just one-shotted Terraria
Fable 5.1 is now registered in the AWS Bedrock API, so the release is probably getting really close
Not sure I’d bet on the 48 hour estimate, but a release within the next week feels pretty realistic.
Average Codex week
Reminder: he’s doing it again today
Claude Code limits are growing by -25% on September 14
Tibo reset Codex limits, and they’re way more generous now
Looks like the limit issue is finally fixed.
Someone ran 52 AI models through the Political Compass
They all turned out to be left-libertarian, except Grok.
Today, August 29, we’re getting a reset
Don’t fumble it like last time and max out your Codex quota before then.
OpenAI added Luna Reserve after you hit your usage limits
The official docs now mention Luna Reserve. Selected Plus and Pro accounts can keep working in Codex and ChatGPT Work with GPT 5.6 Luna after…
Grandpa Tibo reset the limits
Tibo Salamanca about to reset the limits
NVIDIA IS BUYING HUGGING FACE FOR $12.9B
According to The Information, NVIDIA has agreed to acquire Hugging Face. The platform generates around $150M in annual revenue, and the deal would…
GLM 5.3 Flash pushes the Pareto frontier
With an intelligence index around 57.5, the model costs about $0.09 per task and already sits on the Pareto frontier. Until September 9, Z AI offers…
Anthropic may already be A/B testing Fable 5.1
People are checking it like this: regular Fable 5 has a January 2026 knowledge cutoff, while Opus 4.6 came out later. If the model knows those dates…
GLM 5.3 FLASH IS OUT
The first natively multimodal model in the GLM 5 series. 320B total parameters, with just 18B active. It beats GLM 5.2 across most benchmarks and…
QWEN3.8 FLASH NEXT IS OUT
The leaked specs were real: • 125B multimodal MoE • Only 6B active • 262K native context, up to 1M Qwen says it beats Claude Opus 4.6 across coding…
QWEN3.8 FLASH NEXT DROPS TODAY
A ModelScope description that was quickly taken down reportedly revealed: • 125B main parameters • 51B N gram embeddings • Only 6B active per token…
CONFIRMED: OX ALPHA IS A NEW GLM MODEL
Bloomberg got confirmation directly from Z.AI that Ox Alpha is a new iteration of the GLM series. Z.AI also said the name was inspired by the…
Tibo suddenly reset Codex limits after bringing back 5-hour limit
OpenAI brings back 5-hour limit for $20 plan
Pro plan will remain unchanged for another month.
Hit my first 1,000 followers on X
Honestly, I didn’t expect to enjoy it this much or for the platform to feel this much like my place. I’ve met a lot of great people from all over…





























