On August 31, Cognition announced that Claude Fable 5.1 is now available across Devin Cloud, Desktop, and CLI — and that the new model runs 54% cheaper per task than Fable 5, measured at $2.68 versus $5.84 on FrontierCode 1.1 Extended (medium). The company expects users to see “approximately 10–25% savings” in production.
The cost reduction comes from Anthropic’s pricing change for Fable 5.1: cache reads now cost $0.25 per million tokens, 75% below Fable 5’s cache-read rate. For agentic workflows that reload reasoning state across tool calls, the cache-read discount is where the money is.
Walden Yan, Cognition’s Chief Product Officer, stated: “We’re moving our Opus 5 traffic in Devin to Claude Fable 5.1 on launch day… a Fable-class model is finally economical for the workloads we’d kept on Opus, starting with code review.”
What shipped
Three surfaces got Fable 5.1 on August 31:
- Devin Cloud: Fable 5.1 available for selection; Fusion routing upgraded to use Fable 5.1 as lead orchestrator, matching Fable 5.1’s standalone FrontierCode score at $1.43 per task — 47% cheaper than a pure Fable 5.1 harness at $2.68.
- Devin Desktop: Model picker now includes Fable 5.1.
- Devin CLI: Fable 5.1 available via the same model selection flow.
The Fusion upgrade is automatic for Cloud users running in preview. Desktop and CLI users pick the model explicitly.
Cognition also announced a Zero Day Retention (ZDR) exemption for eligible Fable customers while Anthropic rolls out its Enterprise Frontier Safeguards program — data retention behavior that some enterprise buyers require before deploying frontier models.
The sequel to July’s Fable-vs-Opus story
This announcement extends Cognition’s July 13 post, which showed Fable 5 costing less per FrontierCode task than Opus 4.8 despite Fable’s higher token price — Fable’s quality let it finish in fewer attempts. Fable 5.1’s cache-read discount makes the same model cheaper on the same task by cutting the reload tax.
The July post also introduced Fusion, Cognition’s dual-agent routing design (a frontier lead plus a cheaper sidekick) that cut costs 35–60% while holding quality. With Fable 5.1 as the lead, Fusion now delivers those savings on top of Fable 5.1’s own per-run improvements.
What Cognition is moving
Yan’s statement signals a platform shift: Opus 5 workloads moving to Fable 5.1, starting with code review. Opus was the economical frontier pick when Fable 5’s cache behavior made it expensive to run continuously; Fable 5.1 removes that penalty.
For teams buying Devin Cloud by the session or the merge, this changes the calculation. Fable was the quality option with a cost trade-off; Fable 5.1 is now both the quality pick and — according to Cognition’s telemetry on its own benchmark — the cheaper one for agentic work.
Method: Cost figures ($2.68, $5.84, $1.43, 54%, 47%, 10–25% user savings) are from Cognition's August 31 blog post. Cache-read pricing ($0.25/MTok, 75% reduction) and Walden Yan's quote are from Anthropic's Fable 5.1 announcements. FrontierCode 1.1 Extended (medium) is Cognition's internal benchmark. The "economical for Opus workloads" framing is Yan's on the record at Anthropic's site.
References
- Fable 5.1 in Devin and Why It’s Cheaper than Opus 5 — August 31, 2026. Task costs, Fusion upgrade, ZDR exemption.
- Anthropic: Claude Fable and Mythos 5.1 — cache-read pricing ($0.25/MTok, 75% reduction).
- Anthropic: Claude Fable — September 1, 2026. Walden Yan quote on moving Opus 5 traffic to Fable 5.1.
- Making Fable Cheaper Than Opus — July 13, 2026. Earlier Fable 5 vs Opus 4.8 cost comparison.
- Devin Fusion: Frontier Performance at 35% Lower Cost — June 2026. Dual-agent routing architecture.
Editorial from Devin Central — a fan news desk, not Cognition. Devin is a trademark of Cognition.