Muse Spark 1.1 vs. Gemini 3.6 Flash: Score, Cost, and Google's Missing Flagship

Both sit at the top of the mid-tier pack, well below frontier models like Anthropic's Claude Fable 5 (59.9) and OpenAI's GPT-5.6 Sol (58.9) .

Cost Comparison: Muse Spark Is Significantly Cheaper

Muse Spark 1.1 is roughly 1.6× cheaper per output token .

Competitive Context: A High-Profile Exchange

The comparison drew public attention when Meta's highest-paid employee, Alexandr Wang, shared a leaderboard showing Muse Spark 1.1 outperforming all Google models — sparking a high-profile exchange with Google's CEO .

Google's Delayed Gemini 3.5 Pro Flagship

Gemini 3.5 Pro remains unreleased and has missed three consecutive deadlines:

• May 19, 2026 (I/O): CEO Sundar Pichai announced the model onstage and asked the audience to "give us until next month" — implying a June launch .
• June deadline missed: The model only entered a limited Vertex AI enterprise preview; public launch slipped to July .
• July 17 deadline missed: The rebuilt model, which had been restarted with a longer pre-training cycle, still suffers from hallucinations and inconsistent outputs that prevent it from clearing reliability standards . Google abandoned the original base model entirely in favor of a deeper pre-training cycle .
• Current status: As of July 21, 2026 — when Google released the lightweight Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash-Cyber models — the company gave no timing update for Gemini 3.5 Pro. Google says only that it is "being tested with partners and coming soon" . Bloomberg reports the model is "months behind schedule" with coding capabilities falling short of internal goals .
• Market impact: The repeated delays, combined with high-profile researcher departures to OpenAI and Anthropic, erased roughly $225 billion in Alphabet's market value in a single trading session .