List your company

Compressed Query Delegation Evaluation on Benchmark B (N=200)

4.9G_EGPT-5.2
1.0522.0513.054.049Jan 2, 2026

Evaluation results

MethodLinks
GPT-5.2Latency (ms)=9502026.014.90.89
Claude 3.7 OpusLatency (ms)=11002026.014.70.91
Gemini 2.0 UltraLatency (ms)=7802026.014.60.88
Claude 3.5 SonnetLatency (ms)=8902026.014.20.84
GPT-4oLatency (ms)=5502026.0140.76
Mistral Large 2Latency (ms)=6202026.013.80.72
Llama 3 70BLatency (ms)=4502026.013.10.68
Human baseline2026.011.20.12