What We Know
REST OF WORLD:Court backs Catalan amnesty, while Puigdemont’s return remains legally and politically unsettledAI:OpenAI launches GPT-4.1 family with million-token context and software benchmark resultsCYBERSECURITY:South Korea orders exposed-system audits after bank breaches as AI involvement remains disputedGAMING:Team Heretics reportedly plans first Rocket League roster with AtomiK, ajg and DoritoPOLITICS:Delhi detains protesters as election-chief demonstration faces restrictions and conflicting arrest countsTOP STORIES:Riyadh airport reports describe a second attack, with one death and 11 injuriesSPORTS:Rayo Vallecano fans protest club ownership as Athletic Club match is haltedMARKETS:U.S. stocks retreat after record highs as Treasury yields and oil prices riseREST OF WORLD:Court backs Catalan amnesty, while Puigdemont’s return remains legally and politically unsettledAI:OpenAI launches GPT-4.1 family with million-token context and software benchmark resultsCYBERSECURITY:South Korea orders exposed-system audits after bank breaches as AI involvement remains disputedGAMING:Team Heretics reportedly plans first Rocket League roster with AtomiK, ajg and DoritoPOLITICS:Delhi detains protesters as election-chief demonstration faces restrictions and conflicting arrest countsTOP STORIES:Riyadh airport reports describe a second attack, with one death and 11 injuriesSPORTS:Rayo Vallecano fans protest club ownership as Athletic Club match is haltedMARKETS:U.S. stocks retreat after record highs as Treasury yields and oil prices rise
← Back to feed
DevelopingJust now

OpenAI launches GPT-4.1 family with million-token context and software benchmark results

  • 2 sources analyzed
  • Source mix: Web
  • Momentum: Developing

What We Know

OpenAI has shipped GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano in its API as a new model family.Backed by 1 sourcesaiincontext.com All three versions are described as supporting a 1-million-token context window and capabilities spanning text generation, reasoning, and visual reasoning.Backed by 1 sourcessheratonhq.com

The available reporting highlights a 54.6% score for GPT-4.1 on SWE-bench Verified, a benchmark focused on software-engineering tasks.Backed by 1 sourcesaiincontext.com The pricing information surfaced for the family includes GPT-4.1 nano at $0.10, although the supplied material does not specify the pricing unit or provide the full pricing structure.Backed by 1 sourcesaiincontext.com

GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano are described as sharing a 1-million-token context window and text generation, reasoning, and visual reasoning capabilities.Context from one sourcesheratonhq.com

Source Comparison

Aligned reporting
1 corroborates - 1 adds context - 0 conflicts