gpt-6IMAGE
Opus 5.5 vs GPT-6 Sol: measured tokens-per-second on OpenRouter
A post shares OpenRouter TPS figures: Opus 5.5 Fast exceeds 150 tok/s, GPT-6 Sol does 52 standard and 55 fast, and Astra 25, questioning the value of fast-mode surcharges.
aipulsedaily
@aipulseda1lyArchived 2026.09.23Originally posted 2026.09.23

aipulsedaily @aipulseda1lyOriginally posted 2026.09.23
Anthropic completely destroyed OpenAI on this round. The current results and the TPS are insane. Opus 5.5 Fast is doing over 150 tokens per second on OpenRouter. Standard Opus 5.5 does around 70. GPT-6 Sol does 52 on standard and 55 on fast. GPT-6 Astra does 25. Both labs charge double for fast mode. Anthropics is over 2x faster. OpenAIs Sol fast is 6% faster.
x.com/aipulseda1ly/status/2102497522220458070Archived 2026.09.23 10:16