Model comparison
GPT-4 Turbo vs Llama 4 Scout
GPT-4 Turbo is the stronger model overall, scoring 30.5 to 27.7 on the Noometry Index. Llama 4 Scout costs 100× less per token, which makes it the better buy when GPT-4 Turbo's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
MetaSummary
- They share 26 benchmarks with published results for both. GPT-4 Turbo scores higher in 5 categories and Llama 4 Scout in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where GPT-4 Turbo leads 33.8 to 20.2.
- The biggest single-benchmark swing is MATH Level 5: 46.7% for GPT-4 Turbo and 62.3% for Llama 4 Scout.
- Llama 4 Scout is cheaper at $0.10 / $0.30 per million input/output tokens, against $10 / $30 for GPT-4 Turbo.
- Llama 4 Scout has downloadable open weights; the other is API-only.
Side by side
| GPT-4 Turbo | Llama 4 Scout | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 30.5 | 27.7 |
| Released | 2023-11-06 | 2025-04-05 |
| Weights | Proprietary | Open |
| Context window | 128K | 128K |
| Max output | 4K | 4K |
| Input $ / M tokens | $10 | $0.10 |
| Output $ / M tokens | $30 | $0.30 |
| Results tracked | 36 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4 Turbo leads
GPT-4 Turbo: 33.8 (#249), Llama 4 Scout: 20.2 (#339)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| LMArena Coding | 1268 | 1286 |
| BigCodeBench Complete | 58.2% | 43.1% |
| SWE-bench Verified (bash only) | — | 9.1% |
| SciCode | — | 17% |
| WeirdML | 18% | — |
| BigCodeBench Instruct | 48.2% | — |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73.3% | — |
Agentic & Tool Use Not comparable
GPT-4 Turbo: —, Llama 4 Scout: 24.6 (#119)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.1% |
| METR Time Horizons | 36.7% | — |
Reasoning GPT-4 Turbo leads
GPT-4 Turbo: 15.3 (#317), Llama 4 Scout: 9.1 (#345)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1266 |
| DTBench | 61.6% | 57.9% |
| LMCA | 9.8% | 12% |
| Epoch Capabilities Index | 127.25 | 129.64 |
| ForecastBench | 59.4 | 57.5 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | 25.1% | — |
| Kagi LLM Benchmark | — | 36.9% |
| ARC-AGI-1 | — | 0.5% |
| CritPt | — | 0% |
| Chess Puzzles | 6% | — |
Math Llama 4 Scout leads
GPT-4 Turbo: 9.0 (#322), Llama 4 Scout: 19.6 (#286)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 6.7% | 7.8% |
| LMArena Math | 1272 | 1287 |
| MATH Level 5 | 46.7% | 62.3% |
| FrontierMath (Tiers 1-3) | 0.7% | — |
| Omni-MATH | — | 37.3% |
| FrontierMath (Feb 2025 set) | — | 0% |
Knowledge Llama 4 Scout leads
GPT-4 Turbo: 24.3 (#268), Llama 4 Scout: 31.9 (#217)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| GPQA Diamond | 46.6% | 51.8% |
| LMArena Expert | 1223 | 1235 |
| MMLU-Pro | — | 74.2% |
| Confabulations | 28.4% | — |
| Vectara Hallucination Rate | — | 7.7% |
| GPQA (HELM) | — | 50.7% |
| MMLU | 81.3% | — |
Multimodal Llama 4 Scout leads
GPT-4 Turbo: 30.6 (#110), Llama 4 Scout: 32.2 (#102)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| LMArena Vision | 1090 | 1118 |
| SpatialViz-Bench | — | 34.2% |
Multilingual Too close to call
GPT-4 Turbo: 40.5 (#216), Llama 4 Scout: 41.0 (#212)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| LMArena Non-English | 1245 | 1252 |
| LMArena Chinese | 1242 | 1255 |
| LMArena French | 1276 | 1282 |
| LMArena German | 1259 | 1272 |
| LMArena Japanese | 1194 | 1206 |
| LMArena Korean | 1187 | 1207 |
| LMArena Russian | 1259 | 1263 |
| LMArena Spanish | 1260 | 1278 |
Instruction Following Too close to call
GPT-4 Turbo: 65.8 (#216), Llama 4 Scout: 65.8 (#217)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| LMArena Instruction Following | 1249 | 1248 |
| IFEval | — | 81.8% |
Long Context GPT-4 Turbo leads
GPT-4 Turbo: 38.0 (#206), Llama 4 Scout: 27.5 (#294)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| LMArena Longer Query | 1254 | 1265 |
| Fiction.LiveBench | — | 36% |
Writing & Preference GPT-4 Turbo leads
GPT-4 Turbo: 47.7 (#206), Llama 4 Scout: 37.0 (#261)
| Benchmark | GPT-4 Turbo | Llama 4 Scout |
|---|---|---|
| LMArena Text | 1272 | 1279 |
| LMArena Creative Writing | 1269 | 1249 |
| LMArena Multi-Turn | 1267 | 1280 |
| EQ-Bench Creative Writing | — | 783 |
| WildBench | — | 78% |
Frequently asked questions
Is GPT-4 Turbo better than Llama 4 Scout?
GPT-4 Turbo is the stronger model overall, scoring 30.5 to 27.7 on the Noometry Index. Llama 4 Scout costs 100× less per token, which makes it the better buy when GPT-4 Turbo's lead doesn't matter for your workload.
Which is cheaper, GPT-4 Turbo or Llama 4 Scout?
Llama 4 Scout is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; GPT-4 Turbo lists at $10 and $30.
Is GPT-4 Turbo or Llama 4 Scout better for coding?
GPT-4 Turbo scores higher on coding benchmarks: 33.8 versus 20.2 in the Noometry coding category.
Which has the bigger context window?
Both accept 128K tokens.
How many benchmarks do GPT-4 Turbo and Llama 4 Scout share?
26 benchmarks have published results for both models. GPT-4 Turbo has 36 scored results on Noometry and Llama 4 Scout has 43.