Changelog

What's new, improved, and fixed in LaneMind.

v0.20.10

Newest xAI & Google Models: Grok 4.7 & Gemini 3.8 FlashSeptember 29, 2026

Grok 4.7, xAI's newest model, and Gemini 3.8 Flash, Google's newest Flash model, are now available. Both cost the same credits as the models they replace.
If you used Grok 4.5 or Gemini 3.6 Flash, you are moved to Grok 4.7 or Gemini 3.8 Flash automatically.

Grok models now connect through xAI's new model addresses, so they keep working after the old addresses are retired.
Grok and Gemini model descriptions in Settings now show credits per match report and per Ask Coach answer, like the other models.

Newest AI Models: Claude Fable 5.1 & GPT-6September 29, 2026

The newest models from Anthropic and OpenAI are now available: Claude Fable 5.1 (new default), Claude Opus 5.5 and Claude Sonnet 5.5, plus GPT-6 Astra, GPT-6 Sol and GPT-6 Luna.
If you used a previous model, you are moved to its successor automatically. Every successor costs the same or fewer credits — Opus 5.5, GPT-6 Sol and GPT-6 Luna are cheaper than the models they replace.

Model descriptions in Settings now show credits per match report and per Ask Coach answer. Live tips always cost about 1 credit, whichever model you pick.

Using your own OpenAI API key with a GPT-6 model could return empty answers. The app now uses the correct request format for GPT-6.

Replay Safety, Cleaner Item Timelines, STRATZ FixAugust 31, 2026

Watching a replay or spectating a game no longer triggers live AI coaching or match recording. Replays used to burn coaching credits on advice nobody could act on, and could mix someone else's match into your stats.
The item build timeline could list the same purchase twice (for example a doubled Boots of Speed or Wind Lace a minute apart). The tracker sometimes saw an item in two inventory slots at once during partial game updates — for instance right after a courier delivery — and logged it as a second purchase. Purchases are now only recorded from complete inventory snapshots, so the timeline matches what you actually bought.
The STRATZ tab crashed for Divine and Immortal players with a "DIVINE_IMMORTAL is not a valid RankBracket" error. Your rank bracket is now sent to STRATZ correctly, so the tab loads for every rank.

Live Coaching Costs Up To 10x LessAugust 30, 2026

Live coaching tips are now generated by fast, economical AI models regardless of which model you pick in Settings. Premium reasoning models (like Claude Fable 5) spend thousands of hidden "thinking" tokens on every 30-second tip, which could drain a trial in minutes — with this change a full match of live coaching costs a fraction of what it did before, while tips stay just as sharp.
Your chosen premium model is still used where its reasoning actually shows: post-game analysis reports and the in-game Ask Coach chat.
New installs now default to a tip every 60 seconds instead of 30. You can still choose 15 or 30 seconds in Settings if you prefer faster updates.