Reflection AI models, access and benchmark guide
Reflection AI model and access guide with a practical US-English evaluation brief and links to publisher materials.

👁 111 reads
A Beam run may be evaluated as a language-model answer or as part of a tool-using coding setup. For a municipal service note I would measure exact retrieval first, then separately judge whether a code agent edits the correct file.
Choosing the first job
Extract a temporary bus-stop ID from an invented notice and update the matching route fixture with no other changes.
Models in the November 17 snapshot
| Model and track | Placement | Reported score |
|---|---|---|
| Beam 1.0 (AA intelligence) | 38 | 18 index points |
Release record. Publisher documentation identifies the product family. The ranked configurations include Beam 1.0.
Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →Sources
- https://reflection.ai/ (accessed November 17, 2026)
- https://huggingface.co/reflectionai (accessed November 17, 2026)
- https://artificialanalysis.ai/leaderboards/models (accessed November 17, 2026)
More on this from QUANTSLANT
benchmarks·Pruna AI models, access and benchmark guidePruna AI model and access guide with a practical US-English evaluation brief and links to publisher materials.