Gemini 3.5 Flash hits general availability May 29 — Google's agent-default model takes Terminal-Bench 2.1 at 76.2 percent and prices at $1.50/$9 per 1M tokens
Google's Gemini 3.5 Flash reached general availability on May 29, 2026 — pricing at $1.50 input / $9 output per 1M tokens with a 1M-token context window, scoring 76.2% on Terminal-Bench 2.1, beating Gemini 3.1 Pro on coding and agentic benchmarks at 4x the speed of comparable frontier models. The GA release positions 3.5 Flash as the default agent-runtime model across Gemini Enterprise, Antigravity, and Vertex AI — Google's strategic bet that the next AI wave is agents, not chatbots.
The benchmark-and-pricing combination is the substantive piece. 76.2% on Terminal-Bench 2.1 is competitive with the top frontier-model benchmarks on agentic-coding workloads; the 1656 Elo on GDPval-AA and 83.6% on MCP Atlas extend the same finding into the agentic-task and tool-calling axes. The $1.50/$9 per 1M-token pricing is roughly 60% below Anthropic's Opus 4.8 and meaningfully below GPT-5.5 on the comparable-capability tier — Google is undercutting the price floor for frontier-class agent-runtime tokens. The 4x speed-multiplier versus comparable frontier models is the deployment-economics piece: Macquarie Bank piloting 100+ page document reasoning, Shopify running parallel sub-agents for merchant forecasts, Salesforce wiring Agentforce — these are workloads where latency and throughput matter as much as capability.
The strategic positioning is the agent-default thesis. Through 2024-2025 the dominant Gemini-release pattern led with the heaviest Pro-tier flagship model — competitive benchmark posture, premium pricing, demonstrable frontier-leadership claim. The 3.5 launch inverts that pattern by leading with Flash and deferring Pro: the strategic bet is that the agent-runtime workload is the volume-determining workload through the next cycle, and that the pricing-and-speed-and-quality combination of Flash wins more deployment surface than the flagship-Pro headline benchmark would. Anthropic's $30B Series H at near-trillion valuation reflects the alternative bet — premium frontier capability commands premium pricing — and the two strategies will compete for market share through the next several quarters.
Google Blog — Gemini 3.5: frontier intelligence with action → · TechCrunch — With Gemini 3.5 Flash Google bets next AI wave on agents not chatbots → · MarkTechPost — Google Introduces Gemini 3.5 Flash I/O 2026 faster cheaper agents coding →