Gemini 3.8 Flash Surfaces in Shadow Rollout as Leaked Benchmarks and Pricing Circulate

According to the Wall Street Journal, Google plans to release Gemini 3.8 Flash (codename "Skimaki") as early as Wednesday, focused on coding, with insiders claiming major advances in agentic coding; the company has also abandoned the 3.5 Pro checkpoint. As of the posts, Google had not issued an official announcement, but the model has surfaced via shadow rollout in consumer apps, deployment environments, and the Google Cloud Console, with leaked benchmarks circulating widely and Artificial Analysis publishing third-party results.

Confirmed

• WSJ reports Gemini 3.8 Flash brings major agentic coding advances, launching as early as Wednesday (relayed by @kimmonismus, @Hesamation, @xiaohu, etc.)
• In A/B tests on Google's internal coding tool Jetski, Google engineers preferred Gemini 3.8 Flash over Anthropic's Opus model (@kimmonismus citing WSJ)
• Multiple users observed the model in consumer apps (@markk, @IamNotGorbachev WebUI screenshots); @Able-Line2683's Free vs Pro/Ultra account comparison confirms it is being pushed to paid accounts only — a shadow rollout
• @testingcatalog found Gemini 3.8 Flash on a Google Cloud Console quota page, a routine pre-launch step
• @clmt says the model is released with major gains over 3.7 Flash in agentic coding and general knowledge work, and is available to all users on Antigravity (no Google announcement seen at the time)
• WSJ says Google has officially dropped the 3.5 Pro checkpoint due to limited gains over the Flash line

Unconfirmed

• @teortaxesTex relays leaked pricing matching 3.7 Flash: $0.75/Mtok input, $3.75/Mtok output, $0.075/Mtok cached input, with launch in 7 hours — unverified
• @kimmonismus's leaked benchmarks (self-described as unverified) show it beating Opus 5 and "Sol 5.6" on Terminal Bench 2.1 and HLE; @flavioAd expressed amazement and shared analysis; @Able-Line2683's repeated benchmark screenshots cannot be verified
• @ExpensiveSyrup6529 shared Artificial Analysis third-party benchmark results (image gallery), figures not officially confirmed
• @OfficialLoganK posted a 73.7% DeepSWE 1.1 score with eval link; @vedantmisra and @haider1 relayed claims of topping DeepSWE v1.1 at 74% (vs 65% for 3.7 Flash), at the cost of high output token usage; @ChrisGPT cites 89.4% on Terminal-Bench 2.1 while noting that leaderboard is outdated (now at version 4). All third-party, unconfirmed by Google
• Reports relayed by @Hesamation speculate Google may skip 3.5 Pro and go straight to Gemini 4 after training completes — media speculation
• @markk predicted a same-day launch; Andrew Curran (via @TheZachMueller) said "within hours" (the post also mentions Commerce Secretary Lutnick confirming semiconductor tariffs to CNBC) — take WSJ's original report as authoritative

Why it matters

• If a Flash model already beats Anthropic's flagship in internal tests, it reshapes the mid-size model landscape; the leaked Terminal Bench 2.1, HLE, and DeepSWE 73.7%–74% scores would further support this if genuine, though high token usage could raise real-world costs
• Dropping 3.5 Pro suggests a two-tier "lightweight Flash + Gemini 4 flagship" roadmap, skipping intermediate versions
• Rumored pricing identical to 3.7 Flash would mean stronger coding at the same price, pressuring competitors; full availability on Antigravity signals Google is making agentic coding a core distribution scenario

2026-09-02 ~ 2026-09-03 · 45 related posts

Full story(6 episodes)→

• Episode 1: Gemini 3.8 Flash Surfaces in Shadow Rollout as Leaked Benchmarks and Pricing Circulate(2026-09-02, 45 posts)
• Episode 2: Google Launches Gemini 3.8 Flash and Cyber Variant(2026-09-02, 110 posts)
• Episode 3: HF Engineer Praises Gemini 3.8 as Well-Rounded Agentic Model(2026-09-03, 3 posts)
• Episode 4: Gemini Flash ships 3.6-3.8 in six weeks, testers complain of churn(2026-09-03, 2 posts)
• Episode 5: Developer claims Gemini 3.8 Flash is overfit to public benchmarks(2026-09-03, 2 posts)
• Episode 6: Gemini 3.8 Flash Tops Image Reasoning Benchmarks, 30% Faster(2026-09-04, 2 posts)

• [source] WSJ: Gemini 3.8 Flash drops tomorrow, preferred over Opus in internal coding tests — kimmonismus · 2026-09-02
• Rumor: Gemini 3.8 Flash Arrives Tomorrow, Skipping 3.5 Pro — Hesamation · 2026-09-02
• WSJ: Gemini 3.8 Flash to release today, 3.5 Pro checkpoints scrapped — Mysterious_Bed_1804 · 2026-09-02
• Reddit screenshot claims Gemini 3.8 Flash appeared today, unverified — RootDeveloper-DS · 2026-09-02
• Gemini 3.8 Flash hits consumer app ahead of release — brandon_galang · 2026-09-02
• Gemini 3.8 Flash prematurely appears in consumer app ahead of imminent launch — CtrlAltDwayne · 2026-09-02
• [source] Gemini 3.8 Flash spotted rolling out to Pro/Ultra accounts — Able-Line2683 · 2026-09-02
• Screenshots suggest Gemini 2.0 Flash 3.8 is rolling out — Maglcite · 2026-09-02
• Gemini 3.8 Flash Appears to Be Rolling Out to Some Users — ModdingCentral · 2026-09-02
• Rumor: Gemini 3.8 Flash launching today with pricing details revealed — teortaxesTex · 2026-09-02
• Gemini 3.8 Flash release confirmed for today — NBMVegeta · 2026-09-02
• Rumor: Google DeepMind to release Gemini 3.8 Flash today — mark_k · 2026-09-02
• Unreleased Gemini 3.8 Flash spotted selectable inside the Gemini app — 赛博禅心 · 2026-09-02
• Google scraps Gemini 3.5 Pro, to launch Gemini 3.8 Flash tomorrow with coding gains over Opus — 新智元 · 2026-09-02
• Gemini 3.8 Flash spotted on Google Cloud Console, hinting at imminent release — PleasantSir9581 · 2026-09-02
• Gemini 3.8 Flash Expected to Drop Tonight, Echoing WSJ Report — xiaohu · 2026-09-02
• WSJ: Google Set to Ship Coding-First Gemini 3.8 Flash, Codenamed Skimaki — ChuckDBrooks · 2026-09-02
• Gemini 3.8 Flash rumored imminent, but users may need convincing — CtrlAltDwayne · 2026-09-02
• Gemini 3.8 Flash may have stealth-launched on Gemini web, users report — koltregaskes · 2026-09-02
• Leak: Gemini 3.8 Flash expected to drop tonight amid endless minor version bumps — lipeng0820 · 2026-09-02
• Gemini 3.8 Flash spotted hidden in model picker for Italy Pro users — HrmhsMox · 2026-09-02
• Gemini 3.8 Flash rumored to launch within hours; G20 talk says models will match top scientists in 12 months — TheZachMueller · 2026-09-02
• Gemini 3.8 Flash spotted available in WebUI — IamNotGorbachev · 2026-09-02
• Gemini 3.8 Flash spotted in Google deployment, launch possibly hours away — Existing_Hat_1064 · 2026-09-02
• Second Alleged Gemini 3.8 Flash Benchmark Screenshot Circulates on Reddit — Able-Line2683 · 2026-09-02
• Leaked Gemini 3.8 Flash benchmarks claim wins over Opus 5 on Terminal Bench 2.1 — kimmonismus · 2026-09-02
• Leaked Gemini 3.8 Flash benchmarks reportedly beat Opus 5 on Terminal Bench, HLE — kimmonismus · 2026-09-02
• Gemini 3.8 Flash starts showing up for users on Reddit — ModdingCentral · 2026-09-02
• Gemini 3.8 Flash benchmarks reportedly beat Opus 5 and Sol 5.6 on most tests — tkipf · 2026-09-02
• Gemini 3.8 Flash benchmarks drop, and one analyst's reaction is 'holy shit' — flavioAd · 2026-09-02
• Artificial Analysis benchmarks for Gemini 3.8 Flash surface on Reddit — Expensive_Syrup_6529 · 2026-09-02
• Gemini 3.8 Flash scores 73.7% on DeepSWE 1.1 benchmark — OfficialLoganK · 2026-09-02
• Gemini 3.8 Flash scores 73.7% on DeepSWE, beating Fable 5.1 by 6 points — cedric_chee · 2026-09-03
• Gemini 3.8 Flash hits 73.7% on DeepSWE, nearly matching Opus 5 and GPT-5.6 Sol — cedric_chee · 2026-09-03
• Unverified report: Gemini 3.8 Flash out with big agentic coding gains — clmt · 2026-09-03
• Gemini 3.8 Flash posted scoring 73.7% on DeepSWE 1.1 — doodlestein · 2026-09-03
• Unverified claim: Gemini 3.8 Flash tops DeepSWE v1.1 coding benchmark — vedantmisra · 2026-09-03
• Gemini 3.8 Flash reportedly out: 89.4% on Terminal-Bench 2.1, 73.7% on DeepSWE — ChrisGPT · 2026-09-03
• Gemini 3.8 Flash benchmark chart shared on Reddit — NBMVegeta · 2026-09-03
• Gemini 3.8 Flash ties for top of DeepSWE at 74%, but burns 1.34x more output tokens — haider1 · 2026-09-03
• Testing unreleased Gemini 3.8 Flash: no citations shown for top-of-funnel queries — gaganghotra_ · 2026-09-03

4 near-duplicate retellings: freethinker1706 · testingcatalog · Able-Line2683 · daniel_mac8