Compare / Gemma 3 4B vs Llama 4 Scout

Gemma 3 4B vs Llama 4 Scout

Which is cheaper, which is more capable, and which does better at each kind of work, from 29 results on 5 sources that measured both models.

GoogleGemma 3 4B Released 11 Mar 2025 · $0.040 in · $0.050 out
MetaLlama 4 Scout Released 5 Apr 2025 · $0.10 in · $0.30 out
Gemma 3 4B wins 55 too close to callLlama 4 Scout wins 19

Of the 29 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 Meta, every model from both labs.

Gemma 3 4B or Llama 4 Scout? The short answer

They score about the same, so pick Gemma 3 4B: it costs 3.5 times less per token.

  • Gemma 3 4B is cheaper: $0.042 against $0.15 per million tokens, 3.5 times less.
  • Llama 4 Scout scores higher on the Artificial Analysis Intelligence Index: 6.5 against 4.8.
  • Llama 4 Scout is newer: released 5 Apr 2025, against 11 Mar 2025 for Gemma 3 4B.
  • On third-party benchmarks, on the results both models have: Gemma 3 4B does better for sticking to the facts; Llama 4 Scout for agents and tool use, reasoning and knowledge and writing.

Head to head

4.8
Intelligence Index
6.5
$0.042
Price per million tokens, blendedlower is better
$0.15
131,072
Context window, tokens
1,310,720

Every fact side by side

Gemma 3 4BLlama 4 Scout
Price per million tokens$0.040 in · $0.050 out$0.10 in · $0.30 out
Blended (3 in : 1 out)$0.042$0.15
Cheapest host, blended$0.042 (Parasail, BF16)$0.15 (DeepInfra, FP8)
Intelligence Index4.86.5
Spring Prompt overall––
Output speed, tokens a second–132
Context window131,072 tokens1,310,720 tokens
Released11 Mar 20255 Apr 2025 (newer)
WeightsOpenOpen
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 159 146 of 159
Reasoning and knowledgeHard questions across science, maths and general knowledge 181 of 187 165 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 157 of 169 150 of 169
Sticking to the factsSummarising without inventing things, and factual answers 32 of 139 47 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 4BLlama 4 Scout
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 255 · 231 of 387 measured 1,303 1,322
Business, management and finance · TextArena rating, higher is better 208 · 225 of 380 measured 1,330 1,316
Creative writing · TextArena rating, higher is better 243 · 223 of 385 measured 1,275 1,289
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 246 · 232 of 249 measured 4.8 6.5
GPQA Diamond% of questions, higher is better 227 · 186 of 232 measured 29.1% 58.7%
Humanity's Last Exam% of questions, higher is better 194 · 229 of 250 measured 5.3% 3.8%
Berkeley Function Calling Leaderboard (BFCL) V4 reported by Berkeley Function Calling Leaderboard (BFCL) V4 · 3 results
Overall accuracy · BFCL% correct, higher is better 74 · 53 of 81 measured 19.6%bfcl prompt 28.1%bfcl fc
Irrelevance detection · BFCL% correct, higher is better 71 · 75 of 81 measured 53.9%bfcl prompt 44.9%bfcl fc
Memory · BFCL% correct, higher is better 53 · 55 of 81 measured 8.6%bfcl prompt 8.2%bfcl fc
UGI Leaderboard reported by UGI Leaderboard · 3 results
Writing score · UGIscore out of 100, higher is better 185 · 169 of 230 measured 24.0 26.7
Requested-length error · UGI% off the requested word count, lower is better 33 · 99 of 230 measured 9.0% 19.0%
Style adherence · UGIscore from 0 to 1, higher is better 214 · 66 of 230 measured 0.29 0.36
Vectara Hallucination Leaderboard reported by Vectara Hallucination Leaderboard · 2 results
Hallucination rate · Vectara% of summaries, lower is better 28 · 38 of 104 measured 6.4% 7.7%
Answer rate · Vectara% of documents, higher is better 103 · 65 of 104 measured 67.3% 99.0%
Show all 29 results (15 more)
MeasureGemma 3 4BLlama 4 Scout
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Expert prompts · TextArena rating, higher is better 270 · 228 of 338 measured 1,260 1,308
Instruction following · TextArena rating, higher is better 270 · 237 of 387 measured 1,267 1,299
Writing, literature and language · TextArena rating, higher is better 237 · 224 of 386 measured 1,298 1,305
Artificial Analysis reported by Artificial Analysis · 7 results
Terminal-Bench 2.1% of tasks, higher is better 127 · 119 of 131 measured 0.4% 3.7%
τ-bench banking% of tasks, higher is better 121 · 117 of 121 measured 0.4% 3.3%
Artificial Analysis Coding Indexindex score, higher is better 129 · 123 of 130 measured 2.7 8.2
IFBench% of instructions, higher is better 187 · 138 of 195 measured 28.3% 39.5%
Long-context reasoning (AA-LCR)% of questions, higher is better 226 · 198 of 241 measured 6.7% 27.7%
Terminal-Bench Hard% of tasks, higher is better 180 · 175 of 190 measured 0.8% 1.5%
τ²-bench telecom% of tasks, higher is better 187 · 176 of 193 measured 5.0% 15.5%
Berkeley Function Calling Leaderboard (BFCL) V4 reported by Berkeley Function Calling Leaderboard (BFCL) V4 · 5 results
Multi-turn tasks · BFCL% correct, higher is better 75 · 53 of 81 measured 0.4%bfcl prompt 9.0%bfcl fc
Relevance detection · BFCL% correct, higher is better 1 · 1 of 81 measured 100.0%bfcl prompt 100.0%bfcl fc
Single-turn calls (curated) · BFCL% correct, higher is better 74 · 8 of 81 measured 61.1%bfcl prompt 89.4%bfcl fc
Single-turn calls (user-contributed) · BFCL% correct, higher is better 64 · 34 of 81 measured 60.8%bfcl prompt 74.7%bfcl fc
Web search · BFCL% correct, higher is better 56 · 34 of 81 measured 1.0%bfcl prompt 14.5%bfcl fc

Quick answers

Is Gemma 3 4B better than Llama 4 Scout?

They score about the same, so pick Gemma 3 4B: it costs 3.5 times less per token.

Which is cheaper, Gemma 3 4B or Llama 4 Scout?

Gemma 3 4B is cheaper: $0.042 against $0.15 per million tokens, 3.5 times less.

Which is better for agents and tool use, Gemma 3 4B or Llama 4 Scout?

Llama 4 Scout ranks 146 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 Llama 4 Scout?

Llama 4 Scout ranks 165 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 Llama 4 Scout?

Llama 4 Scout ranks 150 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 Llama 4 Scout?

Gemma 3 4B ranks 32 of 139 models on sale for summarising without inventing things, and factual answers, against 47 for Llama 4 Scout.

Which has the bigger context window, Gemma 3 4B or Llama 4 Scout?

Llama 4 Scout: 1,310,720 tokens, against 131,072 for Gemma 3 4B.

Run Gemma 3 4B and Llama 4 Scout 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.