Artificial Analysis rates DeepSeek V4.1 Flash endpoints consistent with reference
Article timeline
The articles in this story. Times are ET.
Artificial AnalysisOfficialArtificial Analysis finds all 22 DeepSeek V4.1 Flash endpoints consistent with referenceAIArtificial Analysis says all 22 DeepSeek V4.1 Flash 1017 endpoints are statistically in line with the reference, scoring 97.5 to 101.4. The 3.9-point range makes it the most consistent model in the Index.
- Artificial AnalysisOfficialArtificial Analysis shares endpoint accuracy results for six AI models
AIArtificial Analysis publishes endpoint accuracy results for GLM-5.3 Flash, Muse Spark Open, GLM-5.3, DeepSeek V4.1 Flash 1017, Kimi K3, and gpt-oss-120b. The full results and the methodology behind the Endpoint Accuracy Index are linked in the post.
- Artificial AnalysisOfficialArtificial Analysis launches Endpoint Accuracy Index v1.1 with expanded model coverage
AIArtificial Analysis launches Endpoint Accuracy Index v1.1, which scores serverless endpoints as a percentage of self-hosted reference accuracy across BFCL-500, HLE-250 and AA-LCR-25. Across 100+ endpoints benchmarked, nearly 1 in 4 score below the reference, and the update adds GLM-5.3 Flash and Muse Spark Open. Two GLM-5.3 Flash endpoints capped at about 32K output tokens score 22% and 29% on HLE-250 against 55% for the reference.
