Google has introduced Gemini 3.8 Flash, the latest addition to its generative AI lineup, pricing the model at $0.75 per million input tokens. The release positions the product for developers and enterprises seeking cost-efficient inference for large workloads, and it is being framed within the broader Gemini family as a specialised, industry-focused variant rather than a general-purpose flagship.
Independent AI commentator Matthew Berman has highlighted the model's performance on targeted benchmarks, noting that Gemini 3.8 Flash achieves a leading score of 61.4 percent on the Harvey Legal Benchmark, a test designed to evaluate large language models on tasks typically encountered in legal practice. According to the commentary cited in the original report, this places the model at the top of comparative rankings for legal-domain evaluation.
ALSO READ | Apple’s First Foldable iPhone Duo Set for Tonight’s Launch with Leaked Color Lineup and Pricing
On Terminal Bench 2.1, a benchmark measuring coding and command-line reasoning ability, the model reportedly posts a score of 89.4 percent. The combination of a relatively low input token price and strong results on these two specialised tests is being presented as evidence that Google is targeting professional services, software development, and other vertical workflows where domain accuracy matters more than broad conversational fluency.
The $0.75 per million input tokens pricing point places Gemini 3.8 Flash firmly in the mid-tier of frontier-model costs, undercutting premium offerings from rival labs while sitting above smaller open-weight models. Google has not, in the material reviewed, disclosed context-window limits, output-token pricing, or regional availability, details that enterprise procurement teams typically require before committing to deployment.
The model's emphasis on vertical benchmarks reflects a wider industry trend in which AI providers are moving away from one-size-fits-all positioning and instead marketing models against narrow, domain-specific leaderboards. For Google, Gemini 3.8 Flash arrives as competition intensifies across the developer ecosystem, with Anthropic, OpenAI, and open-source alternatives all vying for the same pool of API customers.
