AI News — July 6, 2026 | Talking Fingers

Coding agents

[2026-07-03] Anthropic — Claude Code v2.1.200–201 continues hardening the background-by-default subagent workflow, with two visible behavior changes on top of the reliability work. v2.1.200 renames the “default” permission mode to “Manual” across the CLI, --help, VS Code, and JetBrains (--permission-mode manual and "defaultMode": "manual" now accepted), and changes AskUserQuestion dialogs to no longer auto-continue by default — you opt into an idle timeout via /config instead. The bulk of the release is background-agent robustness: sessions no longer silently stop mid-turn after sleep/wake or when reopening a stalled session, agents no longer refuse to start when a crash leaves a stale daemon.lock whose PID the OS reused, daemon handover now judges build recency by an embedded build timestamp (so a reinstalled older build can’t seize the daemon), and subagents cut off by a rate limit before any output return cleanly instead of empty; v2.1.201 stops Sonnet 5 sessions from using the mid-conversation system role for harness reminders, and a run of screen-reader/terminal-rendering fixes rounds it out. It matters because these are the fixes that make the July 1 background-by-default switch dependable — and the permission-mode rename plus the no-auto-continue AskUserQuestion make the tool’s approval posture explicit rather than implicit. (official)

If you run Claude Code unattended, the AskUserQuestion change is the one to check before upgrading: dialogs no longer auto-continue on their own, so a background run that hits a question now blocks until you either answer it or opt into an idle timeout via /config — the reverse of the old behavior, and easy to miss until an overnight run is found parked on a prompt.

For Software Developers: Kick off a long background-subagent run, close the laptop, and reopen it later — under v2.1.200 the session resumes instead of having silently stopped on sleep/wake, and a subagent that hit a rate limit before emitting anything returns cleanly rather than as a blank result you have to re-run to figure out what happened.

AI cost tracking & telemetry

[2026-07-06] OpenAI — workspace agents left free preview today and are now metered in credits. The extended free window that OpenAI set when it made workspace agents generally available across Business, Enterprise, and Edu (on May 22) ended July 6, and as of today agent runs invoked inside ChatGPT are billed in credits. Workspace agents are Codex-powered cloud agents — an evolution of GPTs that keep running after the user logs off, run on schedules, retain memory across projects, and execute multi-step workflows that write code and call connected apps. OpenAI’s rate card puts a typical end-to-end run on GPT-5.5 at 5–25 credits (a worked example of 20K input + 80K cached-input + 5K output tokens lands at ~7.25 credits); runs invoked outside ChatGPT — e.g. an agent replying in a Slack channel — stay in free preview for now. It matters because it converts always-on autonomous agents from a free experiment into a metered line item teams have to budget per run — the same shift to usage-based agent billing that GitHub Copilot made in June, and exactly the per-run spend that Anthropic’s Enterprise cost controls last week were built to govern. (official, rate card)

Because these agents run on schedules and keep going after the user logs off, the metering bites hardest on the unattended runs: a scheduled agent that fires several times a day now bills 5–25 credits each time whether or not anyone acts on the output, so the first budgeting move is auditing which standing agents still need to run at their current cadence rather than waiting for the invoice to surface the ones that don’t.

Watch list

Fable 5 usage terms (July 7) — the promo ends tomorrow: Fable 5 has been included at 50% of weekly usage limits on paid plans through July 7, after which access moves to metered usage credits. Watch where steady-state credit cost lands relative to the old limits (and whether standard Enterprise seats, which get no included allowance, enable credits at all), and whether the new government-coordinated cyber classifier over-blocks legitimate red-team and security work once the promo window closes. (prior coverage)

A day out, the striking part is that no steady-state credit rate has been published at all — so however July 7 lands, teams running Fable 5 get the real number with zero runway to model it, which is itself a signal of how finished the pricing actually is.
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)

GPT-5.6’s GA date matters less than its first independent benchmark: Sol, Terra, and Luna have been measured only by OpenAI on its own CTF cyber suite, so a GA that still ships no outside numbers would leave the models unevaluable regardless of what the calendar says.
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)

A model card alone would be enough to move this: Gemini 3.5 Pro still has no published specs whatsoever for its 2M-token context or Deep Think mode, so even documentation short of GA would be the first evidence those claims are real rather than roadmap.
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)

With three weeks to the cutover, sequencing is the open question — whether the Ruby, TypeScript, and Python SDKs finish tracking the release candidate before July 28 or after, since a spec that finalizes ahead of its own implementations leaves server authors with no clean version to target.