Mistral Large 4 updated weights and lower pricing on AI Gateway
AI Gateway now serves Mistral Large 4 with updated weights and lower token prices; existing users keep the same model ID.
AI Gateway now uses updated weights for **Mistral Large 4** and has lower input, cached-input, and output token prices. Mistral's release announcement attributes improvements in coding efficiency, cybersecurity, spreadsheets, web research, and prompt-injection resistance to the updated weights.
Builders using the model through AI Gateway can keep `mistral/mistral-large-4`; the existing ID now resolves to the updated version. The material also lists Claude Code, Codex, Cursor, and other harnesses as options for setup through the Vercel CLI.
AI Gateway now uses updated weights for **Mistral Large 4** and has lower input, cached-input, and output token prices. Mistral's release announcement attributes improvements in coding efficiency, cybersecurity, spreadsheets, web research, and prompt-injection resistance to the updated weights. Builders using the model through AI Gateway can keep `mistral/mistral-large-4`; the existing ID now resolves to the updated version. The material also lists Claude Code, Codex, Cursor, and other harnesses as options for setup through the Vercel CLI. No specific price changes or comparative performance measurements are included in the material, and the capability improvements are attributed to Mistral's announcement.
The existing Mistral Large 4 ID now points to updated weights at lower listed input, cached-input, and output prices, so users can trial the revised model without changing that ID. Capability improvements are attributed to Mistral; the source gives no price amounts or comparative measurements, leaving workload evaluation necessary before changing model choice.