Google Launches Gemini 3.7 Flash With Coding Upgrades and 50% Price Cut

On August 14, Google officially launched Gemini 3.7 Flash, a "workhorse" model aimed at coding and agentic workflows: DeepSWE score rose to 65.3%, API pricing is half of 3.6 Flash's through year-end, and it is now live on API, AI Studio, and Antigravity. This is Google's third Flash model in three months, once again dragging down the price-performance bar in the mid-tier model market with faster iteration plus lower prices.

Confirmed

• Launch and positioning: Google DeepMind released Gemini 3.7 Flash on August 14 (Thursday), officially billing it as its most intelligent practical workhorse model, optimized for software coding and automated business tasks; per @soumitrashukla9 relaying Demis Hassabis, the new model brings major upgrades to software engineering, web development, and knowledge work
• Benchmark gains: DeepSWE score climbed from 3.6 Flash's 49.0% to 65.3%; @OfficialLoganK and @koraykv frame it as a cumulative leap from 37.0% to 65.3% on DeepSWE v1.1 within three months; FrontierCode keeps improving from 34.4%, with notable gains on AutomationBench and more
• Independent evaluation: @ArtificialAnlys sees significant progress in agentic capability, speed, and cost-efficiency, with speed and cost landing on the Pareto frontier
• Pricing: Through year-end, API prices are half of 3.6 Flash's — $0.75 per million input tokens and $3.75 per million output tokens, an introductory promotional rate
• Capability details: @GoogleDeepMind says coding tasks such as debugging and problem-solving improved markedly over 3.6 Flash, with stronger UI and app design skills; @jackwoth notes fewer agentic loop failures and less manual intervention; @mitsuhiko and others relay that algorithmic optimization drove a significant intelligence boost in about 3 weeks
• Demos and availability: @googledevs demoed a playable 90s retro pixel game generated from a single prompt, showing off multi-step planning and tool calling; the model is live on API, AI Studio, Antigravity, and other entry points
• Not disclosed: Google did not reveal when its flagship Pro model will launch

Why it matters

• Going from 3.5 to 3.7 in three months is a rare cadence, and the Flash series has become Google's high-frequency mainstay product line
• @SmartAIHustle reads the half-price strategy as "the AI model war entering a price-brawl phase," which will directly cut inference costs for developers and agentic apps while squeezing rivals' mid-tier models
• It stakes out the high-demand coding and agentic space with low prices and strong performance, while the flagship Pro's next move still awaits an official announcement

2026-08-14 ~ 2026-08-14 · 58 related posts

Full story(5 episodes)→

• Episode 1: Gemini 3.7 Flash Benchmarks Leak: High Cost-Performance but Instruction-Following Lags(2026-08-13, 17 posts)
• Episode 2: Gemini 3.7 Flash Launches at Half Price; Flagship Pro Rumored to Wait for 4.0(2026-08-13, 10 posts)
• Episode 3: Google Launches Gemini 3.7 Flash With Coding Upgrades and 50% Price Cut(2026-08-14, 58 posts)
• Episode 4: Google's Gemini 3.7 Flash Pricing Strategy Baffles Developers(2026-08-14, 2 posts)
• Episode 5: Devin adds Gemini 3.7 Flash, matching Claude Sonnet 5 at half the cost(2026-08-14, 3 posts)

• Google Launches Gemini 3.7 Flash: Big Jumps in Coding & Agents at Half the Price — koraykv · 2026-08-14
• Google Demos 3-Agent Team Autonomously Training Robotics Models with Gemini 3.7 Flash — koraykv · 2026-08-14
• [source] DeepMind Details Gemini 3.7 Flash: Major Gains in Coding and UI Design — GoogleDeepMind · 2026-08-14
• Gemini 3.7 Flash Hands-on: More Disciplined in Agentic Loops, Fewer Wasted Turns — _philschmid · 2026-08-14
• Google DeepMind Releases Gemini 3.7 Flash — SRSchmidgall · 2026-08-14
• Google Launches Gemini 3.7 Flash for Coding and Agents — GoogleAI · 2026-08-14
• Google Launches Gemini 3.7 Flash with Enhanced Long-Horizon Coding — dair_ai · 2026-08-14
• Google Launches Gemini 3.7 Flash with 50% Price Cut Until EOY — osanseviero · 2026-08-14
• Google Launches Gemini 3.7 Flash at Half the Price of Predecessor — fofrAI · 2026-08-14
• Google Ships Gemini 3.7 Flash: Better Model, Half the Price — heypearlai · 2026-08-14
• Gemini 3.7 Flash Positioned as Enterprise 'Workhorse' Focusing on Coding and Automation — DavidLinthicum · 2026-08-14
• Google releases Gemini 3.7 Flash with coding and agent gains, lower price and faster responses — gerardsans · 2026-08-14
• Google Exec: Gemini 3.7 Flash Offers Fantastic ROI, Leading Performance-per-Cost — melvinjohnsonp · 2026-08-14
• Gemini 3.7 Flash Released: Major Agentic Gains, Hits Pareto Frontier on Speed and Cost — ArtificialAnlys · 2026-08-14
• Google launches Gemini 3.7 Flash: its most intelligent workhorse model — lmthang · 2026-08-14
• Gemini 3.7 Flash Released: Major Upgrades in Coding and Complex Document Processing — shresbm · 2026-08-14
• [source] Google Rolls Out Gemini 3.7 Flash: Generates Playable 90s Games from a Single Prompt — googledevs · 2026-08-14
• DeepMind CEO Confirms Gemini 3.7 Flash is Lightning Fast — RubenEVillegas · 2026-08-14
• Google's Gemini 3.7 Flash Rolls Out in GitHub Copilot — intellectronica · 2026-08-14
• Google Launches Gemini 3.7 Flash with Major Coding Upgrades — rseroter · 2026-08-14
• Gemini 3.7 Flash Launches at Half the Price: Are AI Models in a Price War? — Smart_AI_Hustle · 2026-08-14
• Google's Gemini 3.7 Flash Hinted by Logan, Users Report Blazing Fast Speed — cgarciae88 · 2026-08-14
• Google Officially Announces Gemini 3.7 Flash, Deepens AI for Science Efforts — divy93t · 2026-08-14
• Google Rushes Out Gemini 3.7 Flash With 50% Price Cut Until Year-End — 量子位 · 2026-08-14

26 near-duplicate retellings: OfficialLoganK · _philschmid · OfficialLoganK · cyrus_zei · testingcatalog · OfficialLoganK · soumitrashukla9 · AI_Andrew · GeminiApp · altryne · kalpeshk2011 · tulseedoshi · sunjiao123sun_ · inductionheads · mitsuhiko · ShikharMurty · ammaar · RubenEVillegas · reach_vb · mervenoyann · jackwoth · GeminiApp · jggomezt · APPSO · rohanpaul_ai · sunychoudhary