Google 7/13 Gemini 3.5 Pro mid-July: agent scaffolds + image-edit endpoint, no price change

2 sources

Google on 2026-07-13 publishes a mid-July tier update for Gemini 3.5 Pro emphasizing agent scaffolds (server-side tool-use loop) + a new image-edit-via-namespace endpoint; pro-tier pricing unchanged. Self-reported figures include MMLU-Pro 86.3, GPQA-Diamond 79.1, SWE-Bench Verified 71.2, BFCL v3 68.9, Terminal-Bench 2.1 78.5. Deployment via Gemini API + Vertex AI.

Why now

It lands one week after Google's 07-09 SensorFM + a 07-10 health-AI thread; this is the third Google-frontier item in the run. Creators can carry 'Google's mid-July tier update' as a hook on top of the prior health-AI / agent scaffold framing.

Why it is worth publishing

Best for agent-scaffold creators + image-edit creators. The new image-edit-via-namespace endpoint is the product hook; the server-side tool-use loop is the engineering hook.

Evidence basis

Single official event with no media corroboration today + benchmark headlines + new endpoint. Triple heat without the badge — the third topPick slot is filled by either Meta or ByteDance depending on how you frame the day.

“Google just shipped a Gemini 3.5 Pro update — same pro-tier price, new agent loop, new image-edit endpoint.”

Angle

Frame the Gemini 3.5 Pro mid-July update as a capability-positioning move (agent scaffolds + image-edit-via-namespace endpoint) rather than a price move — pricing stays put. Lead with the new endpoint + the server-side tool-loop, then quote benchmark headlines with the standard 'vendor-supplied, not reproduced' qualifier.

Format

Single-card info-graph + 30-second talking-head follow-up

Demo idea

Single info-graph card: 'Gemini 3.5 Pro mid-July — 3 lines: (1) agent scaffolds (server-side tool-use loop), (2) image-edit-via-namespace endpoint, (3) pro-tier pricing unchanged'. Then a 30-second talking-head: 'why the image-edit endpoint matters for downstream creators, and why I'm not generalizing BFCL v3 to all tool-use tasks'.

Platform notes

API-license (low risk): Gemini 3.5 Pro is consumed via API — license lives with the API contract, not a downloadable-weights license. Do not paraphrase 'paid API' as 'closed source' without distinguishing. When viewers ask 'is this Gemini still free in the consumer app,' pin the link to the Google blog and quote only what's there.

Usable claims

Google on 2026-07-13 published a mid-July tier update for Gemini 3.5 Pro emphasizing agent scaffolds (server-side tool-use loop) + a new image-edit-via-namespace endpoint; self-reported figures include MMLU-Pro 86.3, GPQA-Diamond 79.1, LiveCodeBench v6 75.4, SWE-Bench Verified 71.2, BFCL v3 68.9, Terminal-Bench 2.1 78.5; pro-tier pricing unchanged; deployment via Gemini API + Vertex AI.

Evidence pipeline

From the news

Google 7/13 Gemini 3.5 Pro mid-July: agent scaffolds + new image-edit-via-namespace endpoint

Breakdown

Walks the Gemini 3.5 Pro mid-July update's two named additions (server-side tool-use loop + image-edit-via-namespace endpoint), the unchanged pro-tier pricing, and the API-contract license framing (no downloadable-weights license). Benchmark headlines cited with the standard 'vendor-supplied' qualifier. No media corroboration today — no Independently verified badge on this topic.

Sources

Risks

Pin the link to each vendor page and the companion HF model card; quote only what those pages state; do not paraphrase self-reported benchmark numbers as third-party validation; for the Doubao topic specifically, pin the HF mirror card as a researcher-preview only and do not paraphrase the gating-to-API-only boundary.
Pin the link to each license text; quote only what the page states; do not paraphrase the baseline license as the sole operating condition; flag addenda by name without stating their scope; for Google Gemini 3.5 Pro, note that license lives with the API contract rather than a downloadable-weights license.

Demo ideas

Pair with Meta Llama 4 Behemoth refresh (both US-lab model updates on the same day) — show as 'two US-lab capability updates, one API-only (Gemini) and one community-licensed (Llama 4).'
For long-term Google-frontier coverage, build a 'Google content tier' timeline (07-09 SensorFM / 07-13 Gemini 3.5 Pro update / future Veo 3 update) — one row per anchor, capability + endpoint columns.
If you cover server-side tool-use vs client-side tool-use, anchor it on the Gemini 3.5 Pro mid-July server-side loop — quote only what the Google blog states; run your own comparison if you generalize.