DASHBOARD

results over the last 7 days — who's slacking?

SPEED LEADERBOARD

google/gemini-3.1-flash-lite 452ms
CHEAP
mistral/mistral-medium-latest 513ms
MID
google/gemini-3.1-flash-lite-preview 569ms
mistral/mistral-small-latest 580ms
CHEAP
mistral/mistral-large-latest 653ms
FLAGSHIP
anthropic/claude-haiku-4-5 700ms
CHEAP
openai/gpt-5.4-mini 740ms
CHEAP
google/gemini-3-flash-preview 1205ms
MID
deepseek/deepseek-v4-flash 1407ms
MID
anthropic/claude-sonnet-4-6 1819ms
MID
anthropic/claude-opus-4-7 1845ms
FLAGSHIP
deepseek/deepseek-v4-pro 1959ms
FLAGSHIP
openai/gpt-5.4 2030ms
MID
openai/gpt-5.5 2379ms
FLAGSHIP
google/gemini-3.1-pro-preview 2990ms
FLAGSHIP

7-DAY UPTIME

mistral/mistral-large-latest 100%
FLAGSHIP 5036/5040 probes
mistral/mistral-medium-latest 100%
MID 5040/5040 probes
mistral/mistral-small-latest 100%
CHEAP 5040/5040 probes
google/gemini-3.1-flash-lite 99%
CHEAP 5002/5040 probes
google/gemini-3.1-pro-preview 99%
FLAGSHIP 1662/1680 probes
google/gemini-3-flash-preview 99%
MID 5002/5040 probes
anthropic/claude-haiku-4-5 -
CHEAP
anthropic/claude-opus-4-7 -
FLAGSHIP
anthropic/claude-sonnet-4-6 -
MID
deepseek/deepseek-v4-flash -
MID
deepseek/deepseek-v4-pro -
FLAGSHIP
google/gemini-3.1-flash-lite-preview -
openai/gpt-5.4 -
MID
openai/gpt-5.4-mini -
CHEAP
openai/gpt-5.5 -
FLAGSHIP

WRONG ANSWERS

% of successful probes where the model did NOT return the expected response · lower is better · daily over 7 days

daily averages in your local timezone · only successful probes