SPACEXAI
4/5 workers samplingPOSSIBLE_MODEL_CHANGE · gpt-5
Leaderboard

Ranks with the units still attached

Capability is not speed is not price. Sort the board by the quantity you actually need. Provider-reported numbers never occupy the same column as GBU measurements.

EndpointEvidence
Llama 4 70B64.371.5$0.007155 ms89099.84%LIVE_MEASURED
Llama 4 70B32.371.5$0.0064140 ms19599.75%LIVE_MEASURED
Gemini 2.5 Flash20.279.4$0.013120 ms28099.97%LIVE_MEASURED
Grok 3 Mini16.674.6$0.013180 ms22099.96%LIVE_MEASURED
Llama 4 405B13.578.2$0.01295 ms21099.20%REPRODUCED
GPT-5 Mini8.5078.9$0.020240 ms16099.93%LIVE_MEASURED
Llama 4 405B8.0578.2$0.03090 ms42099.81%LIVE_MEASURED
Llama 4 405B4.7678.2$0.016180 ms8599.20%REPRODUCED
DeepSeek V3.14.5282.4$0.024310 ms7899.41%LIVE_MEASURED
Llama 4 405B3.4378.2$0.036220 ms15599.68%LIVE_MEASURED
Llama 4 405B2.9778.2$0.032280 ms11099.72%LIVE_MEASURED
Qwen3-Max2.5681.6$0.036260 ms9299.70%LIVE_MEASURED
Gemini 2.5 Pro1.9686.8$0.053340 ms14599.95%LIVE_MEASURED
Mistral Large 31.3280.1$0.068350 ms9899.63%LIVE_MEASURED
Grok 40.92587.3$0.095410 ms13899.94%LIVE_MEASURED
Claude 4.1 Sonnet0.72484.1$0.108390 ms10899.90%LIVE_MEASURED
GPT-50.26888.6$0.262520 ms9599.91%LIVE_MEASURED
GPT-50.21788.6$0.292610 ms8899.88%LIVE_MEASURED
Grok 4 Heavy0.19889.8$0.205890 ms6299.91%LIVE_MEASURED
Claude 4.1 Opus0.10487.9$0.519680 ms7299.86%LIVE_MEASURED
o4-pro0.03390.4$0.7382.40 s4899.74%LIVE_MEASURED