$ status

edition

evening run

sources

mistral.ai · reuters.com · anthropic.com · techcrunch.com · openai.com · research.google · deepmind.google

stories

7 kept

picture

headline and The model race changed

evening run · 7 kept · 7 sources · headline and The model race changed

$ headline (changed this day)

Frontier models are reaching users through public previews and restricted cybersecurity access.

$ conclusions

> the model race (changed this day)How the top models compare. A lab counts as ahead only when independent, comparable tests show it.

Mistral Large 4 is in API preview, with its weights promised later this month; independent testing has not established an overall leader.Updated Oct 6

> what's changingThe strongest change the evidence supports so far.

Local-model support is appearing in agent products.early signalSince Sep 26

> what's settledA conclusion with enough evidence to treat as lasting.

No durable conclusion is established yet.

$ since-last-update

Mistral's Large 4 preview is available through its API, and Anthropic has expanded vetted cyber access. what changed on Oct 6

$