Google releases Gemini 3.6 Flash and two companion models, but benchmarks show it trailing rivals at similar price points
Google launched Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber on July 21; the flagship 3.6 Flash is priced at US$1.50/US$7.50 per million tokens and claims 17% fewer output tokens than its predecessor, but third-party benchmarks show it below Meta Spark 1.1, Grok 4.5, and GPT-5.6 Luna at the same tier
리스트에 추가
아직 리스트가 없습니다.
Summary
Google DeepMind released three new Gemini Flash-tier models on July 21: Gemini 3.6 Flash (general-purpose coding and knowledge work), Gemini 3.5 Flash-Lite (high-volume, low-cost), and Gemini 3.5 Flash Cyber (cybersecurity-tuned). Gemini 3.6 Flash is priced at US$1.50 per million input tokens and US$7.50 per million output tokens. Google claims it produces 17% fewer output tokens than Gemini 3.5 Flash in the Artificial Analysis benchmark, reducing costs in token-heavy workloads. Third-party benchmarks are mixed to negative: WCCFTech reported it trails Meta Spark 1.1, GLM-5.2, GPT-5.6 Luna, Grok 4.5, and GPT-5.6 Terra on intelligence scores at similar price points. The release came while Google's promised flagship model, Gemini 3.5 Pro, which was expected in June, remains undelivered.
The split
WCCFTech leads with the negative benchmark reading, framing the release as Google's weakest model relative to the competition. Fello AI takes a more neutral developer-tool framing, focusing on pricing and whether users should switch. Unite.AI leads with the contrast between three Flash-tier releases and the continued absence of Gemini 3.5 Pro. Gigazine, a Japanese tech outlet, provides the most detailed benchmark breakdown, including the Artificial Analysis index score of 50 and the token output speed of 303.6 tokens per second. Missing from the feed: any Chinese-language AI press or official Google blog post (which would be the primary source for Google's own claims).
By the numbers
- US$1.50 / US$7.50, Gemini 3.6 Flash input / output price per million tokens
- 17%, claimed reduction in output tokens versus Gemini 3.5 Flash (per Google, via Artificial Analysis index)
- 50, Gemini 3.6 Flash's score on the Artificial Analysis index, versus a 31-point median for other models in the same price range
- 303.6, output tokens per second for Gemini 3.6 Flash, versus a 78.5-token median for competing models in the same price range
Why it matters
Google's Flash tier is the most widely-used tier of Gemini in production developer applications. Falling behind Meta Spark, GPT-5.6 Luna and Grok 4.5 in intelligence benchmarks at a comparable price point would accelerate the migration of cost-sensitive workloads away from Google. The delayed flagship (Gemini 3.5 Pro) adds pressure on Google to compete at the top tier, where it has not shipped a new model since Gemini 3.5.
What to watch
- Gemini 3.5 Pro release timeline and benchmark positioning against GPT-5.6 Sol and Fable 5
- Developer adoption rate of Gemini 3.6 Flash versus the previous Flash generation
- Whether Gemini 3.5 Flash Cyber gains traction in enterprise security tooling
- Updated Artificial Analysis composite scores as more third-party tests arrive