Back to rankings

Mistral Vibe: /status Now Discounts Cached Tokens From Session Cost

Mistral VibeView original changelog

Mistral Vibe v2.23.3 updated the /status command to show provider prompt-cache usage and to discount cached tokens from the reported session cost. Previously, cached tokens — which providers typically bill at a reduced rate or not at all — were still counted at full price in Mistral Vibe's cost estimate, inflating the number shown to users. The change gives developers a more accurate, trustworthy picture of what a session is actually costing them.

Key Takeaways

  • /status now shows provider prompt-cache usage directly alongside a session's other stats.
  • Cached tokens are now discounted from the reported session cost instead of being billed at full price in the estimate.
  • The change corrects a previously inflated cost figure that ignored cache discounts entirely.
  • It builds on v2.23.2, which first began tracking cache-hit token counts in session stats.
  • The fix particularly benefits developers running long sessions or workflows that reuse a large, stable context.
  • Shipped in v2.23.3, released August 3, 2026.

More Accurate Cost Reporting

Mistral Vibe v2.23.3 updated the /status command to display provider prompt-cache usage and to discount cached tokens from the session's reported cost. Most model providers charge a reduced rate — or nothing at all — for tokens served from a prompt cache rather than freshly processed. Until this release, however, Mistral Vibe's session cost estimate did not account for that discount, meaning /status could overstate how much a session actually cost, sometimes significantly, in workflows that reuse a lot of context.

Building on Recent Cache Visibility Work

This follows closely on the heels of Mistral Vibe's v2.23.2 release, which began tracking provider cache-hit token counts in session stats for the first time. That earlier change surfaced the raw numbers; this release goes a step further by actually factoring the cache discount into the cost figure users see, rather than just reporting cache hits alongside an unadjusted total.

Why It Matters

For developers running long sessions or repeatedly working against a large, stable context — a common pattern in agentic coding workflows — prompt caching can meaningfully reduce real spend. Having /status reflect that accurately means the cost figure Mistral Vibe reports is something users can actually trust and budget against, rather than a worst-case estimate that ignores caching entirely.