neomanex
Gemini 3.5 Pro logo

Gemini 3.5 Pro

Google · Released May 2026

View pricingCompare models
pending

Google's pre-GA frontier model — now on its fourth postponement and 'months behind schedule' per Bloomberg. The 2M-token context, native multimodal pipeline, and Deep Think reasoning remain real capabilities, but no confirmed GA timeline exists. Coding capability is the core shortfall.

Is it right for you?

Good for

  • Massive context workloads requiring 1M-2M tokens — whole codebase refactors, multi-document legal review, long-horizon research. 2M context is the largest of any production frontier model.
  • Multimodal workflows combining text, images, audio, and video in a single request — Gemini's native multimodal pipeline is the strongest in the market.
  • Google Workspace and Vertex AI ecosystem users — zero migration cost, native integration with Gmail, Docs, Sheets, and enterprise agent platform.

Not good for

  • Autonomous coding agents — no official SWE-bench Pro or coding benchmarks published. Claude Fable 5 leads coding at 80.3% SWE-bench Pro. Wait for independent evaluations before deploying for agentic coding.
  • Production deployments requiring stable, documented pricing — official API pricing has not been published. Third-party estimates range widely ($5-$15 input, $30-$60 output).

Pricing

Input

Not yet published

Output

Not yet published

Context

Est. $5-15/$30-60 per 1M. 2M ctx. Not GA.

View full pricing

Verdict history

Jul 17, 2026
Expected July 2026 -> Delayed — months behind, no confirmed timeline — Bloomberg reports coding capabilities fell short of internal goals; fourth postponement since May 2026 I/O announcement. No new GA date.
Jul 16, 2026
copy edit to brevity standard, detail preserved in children — Density rewrite: compressed 167w/7s to ~80w/3s per the 55–90w verdict_summary standard. The 2M-context, multimodal, and Deep Think facts kept in-summary because benchmarks[] and task_strengths[] are empty on this model; remaining detail (ecosystem, delay narrative, alternatives) is preserved in strengths[], verdict_audiences[], and sources[].

Sources

Verification log

  • Pricing— No changes

    Automated agent

  • Pricing— No changes

    Automated agent

  • Status— Updated

    Automated agent

    Summary re-dated to Jul 7 2026: states GA slipped from June to July (per our own post), still pre-GA with no published pricing or independent benchmarks. Verdict stays pending; no sourced GA announcement with pricing found on WebSearch, so no rating resolution. Explicitly states what is awaited (GA + pricing + benchmarks).

  • Pricing— No changes

    Automated agent

  • Pricing— No changes

    Automated agent

    Pricing unchanged: Not yet published. Pre-GA model delayed to July 2026. Not listed on Google AI pricing page.

  • Pricing— No changes

    Automated agent

  • Pricing— No changes

    Imported at launch

  • Profile— No changes

    Imported at launch

How we evaluate