Breaking

Anthropic Reverses Course on Claude Fable 5, Trimming Usage Limits as GPT-5.6 Sol Rivalry Intensifies

July 19, 2026 at 01:40 EDT

Foundation Models
Sales & Business

Anthropic has walked back an earlier plan to strip its flagship Claude Fable 5 model from subscription plans, instead redeploying it with tighter usage limits in a move widely read as a response to mounting competitive pressure from OpenAI's GPT-5.6 Sol.

June 30, 2026 · Anthropic

Anthropic keeps Claude Fable 5 — but tightens the leash

After nearly pulling its flagship model from subscriptions, Anthropic redeploys Fable 5 with stricter caps — a move widely read as a defensive response to OpenAI's aggressively priced GPT-5.6 Sol.

50%

Weekly usage cap for Fable 5 — now only on Max & Team Premium

−33%

Overall usage limits cut as the bonus period wound down

$100

One-time credit for Pro & Team Standard, pushed to metered usage

Output token price — Fable 5 vs GPT-5.6 Sol

Per million output tokens, once subscription allotments are exhausted

$50

Fable 5

$30

GPT-5.6 Sol

Input tokens follow the same gap: $10 (Fable 5) vs $5 (Sol) per million. US-only inference adds a 1.1× premium.

The competitor setting the bar

GPT-5.6 Sol's benchmark pressure

Coding Agent Index score

73.5%

ExploitBench cybersecurity

Token context window

From full inclusion to tiered access

Originally slated for full subscription inclusion

→

Compute crunch + export pause (Jun 12–30)

→

Redeployed, then narrowed to premium tiers (Jul 20)

Welcomed by Max users

Permanent inclusion justifies the higher tier
Praised for agentic strength on complex coding

Frustration from Pro users

Shifting plans caused confusion; $100 burns fast
Limits feel unpredictable — some weigh switching providers

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.