Gemini 1.0 Pro

// Overview

Gemini 1.0 Pro was Google's mid-tier model in the first Gemini family, announced on December 6, 2023 and made available to developers on December 13, 2023 through Google AI Studio and the Vertex AI Gemini API. Positioned between the lightweight Gemini Nano and the flagship Gemini Ultra, Gemini 1.0 Pro was pitched as the best model for scaling across a wide range of everyday text, reasoning, and coding tasks. A fine-tuned version of Gemini Pro also powered the largest upgrade to Google's Bard chatbot since its launch.

Unlike the text-only assistants of its era, the Gemini line was designed to be natively multimodal. The base gemini-1.0-pro endpoint handled text in and text out, while a companion gemini-1.0-pro-vision variant accepted image input alongside text. The model shipped with roughly a 32K-token context window and an 8,192-token maximum output, modest by today's standards but competitive at launch. Google described Gemini Pro as outperforming GPT-3.5-class models on most academic benchmarks.

Gemini 1.0 Pro has since been retired. Google moved developers onto the far more capable Gemini 1.5 (with its million-token context) and later the Gemini 2.x and 3.x generations, and the 1.0 Pro endpoints were marked unsupported/discontinued across Google AI Studio and Vertex AI in 2025. The model remains historically important as the version that first put the Gemini brand in front of developers and Bard users.

// Benchmarks

1. MMLU (5-shot)71.8%
2. GSM8K (grade-school math)86.5%
3. MATH (4-shot)32.6%
4. HumanEval (code)67.7%
5. Natural2Code69.6%
6. BIG-Bench-Hard75%

Scores on a 0–100 scale (25-point gridlines); higher is better. Each benchmark links to its published source.

// Pricing

Paid Gemini API token pricing for the gemini-1.0-pro text model; model now retired.

Pricing source ↗

// Strengths

• Natively multimodal design — the Gemini 1.0 Pro Vision variant accepted images alongside text
• Strong grade-school math reasoning for its tier (86.5% on GSM8K in the Gemini technical report)
• Broad general-knowledge and reasoning coverage (71.8% MMLU), beating GPT-3.5-class models at launch
• Available day one through both Google AI Studio and enterprise-grade Vertex AI
• Low, predictable token pricing that made it cheap to scale across high-volume text tasks

// Best for

• Multi-turn text and code chat assistants
• Code generation and explanation in popular programming languages
• Summarization, classification, and rewriting over natural-language text
• Image understanding and visual Q&A (via the Gemini 1.0 Pro Vision variant)
• Powering Bard-style conversational reasoning and planning features

// How to access

// Gemini Pro — every version

The full lineage of the Gemini Pro line, newest first. Every version has its own page — click any to compare specs, benchmarks and pricing.