Compare / Gemma 3 4B vs GPT-4o mini (2024-07-18)

Gemma 3 4B vs GPT-4o mini (2024-07-18)

Which is cheaper, which is more capable, and which does better at each kind of work, from 13 results on 2 sources that measured both models.

GoogleGemma 3 4B Released 11 Mar 2025 · $0.040 in · $0.050 out
OpenAIGPT-4o mini (2024-07-18) Released 18 Jul 2024 · $0.15 in · $0.60 out
Gemma 3 4B wins 14 too close to callGPT-4o mini (2024-07-18) wins 8

Of the 13 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.

Results as of 16 November 2026 (catalogue release 2026-11-16-49bdfa045dee). See also Google vs OpenAI, every model from both labs.

Gemma 3 4B or GPT-4o mini (2024-07-18)? The short answer

They score about the same, so pick Gemma 3 4B: it costs 6.2 times less per token.

  • Gemma 3 4B is cheaper: $0.042 against $0.26 per million tokens, 6.2 times less.
  • GPT-4o mini (2024-07-18) scores higher on the Artificial Analysis Intelligence Index: 6.7 against 4.8.
  • Gemma 3 4B is newer: released 11 Mar 2025, against 18 Jul 2024 for GPT-4o mini (2024-07-18).
  • On third-party benchmarks, on the results both models have: GPT-4o mini (2024-07-18) does better for agents and tool use, reasoning and knowledge and writing.

Head to head

4.8
Intelligence Index
6.7
$0.042
Price per million tokens, blendedlower is better
$0.26
131,072
Context window, tokens
128,000

Every fact side by side

Gemma 3 4BGPT-4o mini (2024-07-18)
Price per million tokens$0.040 in · $0.050 out$0.15 in · $0.60 out
Blended (3 in : 1 out)$0.042$0.26
Cheapest host, blended$0.042 (Parasail, BF16)–
Intelligence Index4.86.7
Spring Prompt overall––
Output speed, tokens a second––
Context window131,072 tokens128,000 tokens
Released11 Mar 2025 (newer)18 Jul 2024
WeightsOpenClosed
Developer based inthe USthe US

Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.

Which is better for what

Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.

Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 158 of 159 155 of 159
Reasoning and knowledgeHard questions across science, maths and general knowledge 181 of 187 176 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 157 of 169 149 of 169
Sticking to the factsSummarising without inventing things, and factual answers 32 of 139 135 of 139

Every shared result

Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.

MeasureGemma 3 4BGPT-4o mini (2024-07-18)
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 255 · 236 of 387 measured 1,303 1,318
Business, management and finance · TextArena rating, higher is better 208 · 234 of 380 measured 1,330 1,311
Creative writing · TextArena rating, higher is better 243 · 216 of 385 measured 1,275 1,294
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 246 · 225 of 249 measured 4.8 6.7
GPQA Diamond% of questions, higher is better 227 · 217 of 232 measured 29.1% 42.6%
Humanity's Last Exam% of questions, higher is better 194 · 221 of 250 measured 5.3% 4.2%
Show all 13 results (7 more)
MeasureGemma 3 4BGPT-4o mini (2024-07-18)
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Expert prompts · TextArena rating, higher is better 270 · 250 of 338 measured 1,260 1,285
Instruction following · TextArena rating, higher is better 270 · 243 of 387 measured 1,267 1,293
Writing, literature and language · TextArena rating, higher is better 237 · 207 of 386 measured 1,298 1,314
Artificial Analysis reported by Artificial Analysis · 4 results
Terminal-Bench 2.1% of tasks, higher is better 127 · 110 of 131 measured 0.4% 5.6%
τ-bench banking% of tasks, higher is better 121 · 118 of 121 measured 0.4% 2.9%
Artificial Analysis Coding Indexindex score, higher is better 129 · 118 of 130 measured 2.7 11.4
IFBench% of instructions, higher is better 187 · 179 of 195 measured 28.3% 31.0%

Quick answers

Is Gemma 3 4B better than GPT-4o mini (2024-07-18)?

They score about the same, so pick Gemma 3 4B: it costs 6.2 times less per token.

Which is cheaper, Gemma 3 4B or GPT-4o mini (2024-07-18)?

Gemma 3 4B is cheaper: $0.042 against $0.26 per million tokens, 6.2 times less.

Which is better for agents and tool use, Gemma 3 4B or GPT-4o mini (2024-07-18)?

GPT-4o mini (2024-07-18) ranks 155 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 158 for Gemma 3 4B.

Which is better for reasoning and knowledge, Gemma 3 4B or GPT-4o mini (2024-07-18)?

GPT-4o mini (2024-07-18) ranks 176 of 187 models on sale for hard questions across science, maths and general knowledge, against 181 for Gemma 3 4B.

Which is better for writing, Gemma 3 4B or GPT-4o mini (2024-07-18)?

GPT-4o mini (2024-07-18) ranks 149 of 169 models on sale for what people prefer in blind comparisons, and judged writing quality, against 157 for Gemma 3 4B.

Which is better for sticking to the facts, Gemma 3 4B or GPT-4o mini (2024-07-18)?

Gemma 3 4B ranks 32 of 139 models on sale for summarising without inventing things, and factual answers, against 135 for GPT-4o mini (2024-07-18).

Which has the bigger context window, Gemma 3 4B or GPT-4o mini (2024-07-18)?

Gemma 3 4B: 131,072 tokens, against 128,000 for GPT-4o mini (2024-07-18).

Run Gemma 3 4B and GPT-4o mini (2024-07-18) on your own prompt.

Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.