AI News — July 7, 2026

AI cost tracking & telemetry

• [2026-07-07] Anthropic — Fable 5’s promotional access ended today; the model is now billed through metered usage credits. Since its July 1 global return, Fable 5 had been included at up to 50% of weekly usage limits on Pro, Max, Team, and select Enterprise plans through July 7; as of today that allowance expires and continued Fable 5 use moves to usage credits — the metered top-up mechanism paid Claude plans use once their included limits are exhausted. The underlying API rate is $10 / $50 per million input/output tokens, roughly double Claude Opus 4.8 and the most expensive rate Anthropic has published for a generally available model. In practice, teams that rewired production onto Fable 5 during the promo now meter every call against a per-token budget rather than a weekly cap — and without usage credits enabled in the Console, Fable 5 access simply stops when the included allowance runs out rather than degrading gracefully. It matters because it closes the last open variable from the export-control saga — the actual steady-state cost of the model that drove a month of briefings — and sets an unusually high floor for frontier-tier spend at the same moment OpenAI (workspace agents, July 6) and GitHub (Copilot credit pools) are converting always-on AI from flat plans to per-run metering. (official, rate, usage credits)

  Doubling the per-token rate to $10/$50 is the part that bites quietly: any workflow that swapped Opus 4.8 for Fable 5 while it was free now costs twice as much per call for the same volume, so the first steady-state move is re-checking which of those workloads actually needed the top tier versus which rode the promo simply because it cost nothing.

  For ML / Data Engineers: If a batch or scheduled pipeline got pointed at Fable 5 during the promo, today changes its failure mode — with the weekly-limit cushion gone, the pipeline hard-stops mid-run when the included allowance is exhausted unless usage credits are enabled in the Console, so either turn credits on before the next run fires or pin the stages that don’t need frontier reasoning back to Opus 4.8 at half the per-token cost.

Coding agents

• [2026-07-06] Anthropic — Claude Code v2.1.202 is a routine stability release, with two changes worth noting on top of the fixes. /review <pr> reverts to a fast single-pass review (multi-agent review now lives at /code-review <level> <pr#>), and workflow-spawned agents emit new workflow.run_id and workflow.name OpenTelemetry attributes, so work fanned out across background agents can be traced and cost-attributed per workflow rather than as an undifferentiated pool. A new /config “Dynamic workflow size” setting caps how many agents a dynamic workflow spins up, alongside the usual batch of session-reliability and terminal-rendering fixes. (official)

  Anyone with /review <pr> wired into a script or muscle memory should note the swap: the bare command is the fast single-pass again, and the deeper multi-agent review now runs only when you spell out /code-review <level> <pr#> — so an unchanged habit quietly gets you a shallower review than it did last week.

  For Solution Architects: If you run Claude Code across a team and pipe its telemetry into an observability backend, the new workflow.run_id and workflow.name OpenTelemetry attributes let you attribute background-agent token spend to the specific workflow that fanned it out rather than reading it as one undifferentiated pool — and the /config “Dynamic workflow size” cap gives you a ceiling on how many agents a single workflow can spin up, so a runaway fan-out can’t silently run up the bill.

Watch list

• Fable 5 cyber classifier & enterprise credit access — with the pricing question now answered ($10/$50 via usage credits), the open residuals are behavioral: whether the government-coordinated cybersecurity classifier Fable 5 ships behind over-blocks legitimate red-team and security work now that the model is in steady-state paid use, and whether standard Enterprise seats — which carry no included Fable 5 allowance — can enable usage credits at all. (official, prior coverage)

  Of the two residuals, the enterprise-seat question is the one with immediate money attached: a standard Enterprise seat carries no included Fable 5 allowance, so if those seats can’t turn on usage credits at all, Fable 5 is effectively unavailable to them today regardless of the published rate.

• GPT-5.6 general availability — Sol, Terra, and Luna remain a ~20-partner government-coordinated limited preview; OpenAI still says broad GA is “in the coming weeks,” with reporting pointing to mid-to-late July. Watch for the GA date, whether the vetted-partner gate persists, and the first third-party benchmarks beyond OpenAI’s internal CTF cyber scores. (prior coverage)

  With Fable 5’s pricing now settled, GPT-5.6 is the last frontier model whose real cost and access terms remain entirely unknown — so its GA will be as much a pricing-and-availability reveal as a capability one, worth watching the rate card as closely as the benchmarks.

• Gemini 3.5 Pro (GA) — still a limited Vertex AI enterprise preview with no model card, public benchmarks, or pricing for the promised 2M-token context and Deep Think mode; Google’s only guidance remains a July target. Watch for whether even a model card lands this month or GA slips into Q3. (prior coverage)

  Set against Fable 5’s now-settled pricing and GPT-5.6’s at-least-internal CTF numbers, Gemini 3.5 Pro stands out as the only major frontier preview with nothing public at all — no specs, benchmarks, or price — which makes continued July silence read less like a model that’s merely unpriced and more like one that isn’t finished.

• MCP spec finalization (July 28) — the 2026-07-28 release candidate (stateless core, Extensions framework, Tasks, MCP Apps, authorization hardening, formal deprecation policy) is in its validation window with the Ruby/TypeScript/Python SDKs updating against it. Watch for the remaining SDK support landing — and any breaking issues in Tasks, MCP Apps, or the new MCP-specific HTTP headers — before the cutover. (official, prior coverage)

  For anyone maintaining an MCP server, the practical move before July 28 is to test against the release candidate now rather than waiting for the final spec — the RC is where breaking issues in Tasks, MCP Apps, and the new HTTP headers surface first, and finding them during the validation window is the difference between a routine upgrade and an emergency one after the cutover.