June 2026 archive
June's four releases spanned a new Anthropic price tier and a scheduled retirement: Claude Fable 5 launched as the first 'Mythos-class' model on 2026-06-09, Claude Sonnet 4 and Opus 4 retired from the API on 2026-06-15 on Anthropic's published schedule, GLM-5.2 shipped MIT-licensed open weights on 2026-06-16, and Claude Sonnet 5 closed the month on 2026-06-30 at $2 in / $10 out — the cheapest of the four Claude models tracked on this site.
Every AI model release in June 2026
- 2026-06-30Release
Claude Sonnet 5 released
$2/$10 and 1M context at the family's lowest price — and the intro pricing was later made permanent. Six independent scores on our board; sits well below Opus 5/Fable 5 on coding benchmarks, as priced.
Official source → - 2026-06-16Release
GLM-5.2 reaches general release
MIT-licensed open weights — the most permissive license in this batch of releases; several competitors use bespoke terms that require a commercial agreement above certain revenue or usage thresholds.
Since then: GLM-5.2 was superseded by GLM-5.3 once GLM-5.3's API went live on 2026-08-19, at an unchanged $1.40/$4.40. It remains the only model in the GLM-5 family with independent benchmark coverage on this site.
Official source → - 2026-06-15Deprecation
Claude Sonnet 4 and Opus 4 retired from the API
Announced April 14, gone June 15 — Anthropic's ~60-day retirement cadence in action. If you pin old model versions, this is the clock you're on.
Official source → - 2026-06-09Release
Claude Fable 5 — first Mythos-class model goes GA
A new price tier ($10/$50), not an Opus successor. Independently #1 on Terminal-Bench 2.1 (83.8) — one of the few launch claims a third party has confirmed.
Official source → - 2026-06-01Release
MiniMax M3 — open-weight value play, added to tracker retroactively
Logged 2026-08-26 alongside this site's first MiniMax M3 model page, not on its original launch day. Vendor's Terminal-Bench 2.1 self-report (66.0%) runs 12.4 points above vals.ai's independent bash-only harness (53.56%) — no third source exists to arbitrate. Four other tracked benchmarks (GPQA Diamond, SWE-bench Verified, LiveCodeBench, LiveBench) are independently confirmed. ~205K Hugging Face downloads in the past month.
Official source →