Unverified50% confidenceBenchmarkExact time
After testing OpenAI's Astra model, the Hebbia team saw a double-digit performance improvement on the citation source selection task compared with the other models they tested.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/27/2026
First Seen
Valid until: 1/25/2027
Sources
Related Entities
Related Claims
UnverifiedOpen-source models' performance on benchmarks has increasingly approached, and in some tasks even surpassed, that of closed-source models.65% similarUnverifiedThe model trained using Edu-QuRating-filtered mixed sets achieved a higher overall observed accuracy than the FineWeb-Edu baseline model across nine benchmark tests, with performance gains concentrated on specific tasks.65% similarUnverifiedIn recent years, institutions such as OpenAI have continually publicized their models' progress on mathematical benchmarks, with some models delivering impressive performance on competition-level problems.63% similarUnverifiedOpenAI has a 'Performance Factory' that runs performance tests in the production environment and feeds issues back to Codex once regressions or errors are detected.63% similarUnverifiedWhen optimizing a workload running on thousands of GPUs, Astra found a new optimization through testing and exploring existing experiments that improved speed by 3.3%.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1101638API
curl https://kongchang.com/api/v1/knowledge/claims/1101638MCP
get_claim(id=1101638)