kimi-k3.1-pro and kimi-k3.1-pro-highspeed passed basic text, streaming, JSON, tool round-trip, and image-understanding tests.
Standard input/output rates are $2/$10 and $4/$20 per million tokens, respectively, matching official pricing. Both models have a configured cache-read rate of $0.20. Cache read/write billing and one-hour retention remain unverified in this test round; see the integration overview.
Eligible top-up promotions add 20% bonus credit, bringing effective costs to about 83.3% of list price. Rates may change with official pricing; the Model Pricing tab is authoritative.
← Back to Live Updates · 📚 Browse the archive