Model comparison
GPT-5 Nano vs Llama 4 Scout
GPT-5 Nano is the stronger model overall, scoring 33.5 to 27.7 on the Noometry Index.
Last verified . 38 shared benchmarks.
MetaSummary
- They share 38 benchmarks with published results for both. GPT-5 Nano scores higher in 9 categories and Llama 4 Scout in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in coding, where GPT-5 Nano leads 33.6 to 20.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 81.1% for GPT-5 Nano and 7.8% for Llama 4 Scout.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $0.10 / $0.30 for Llama 4 Scout.
- GPT-5 Nano accepts more context: 400K tokens versus 128K.
- Llama 4 Scout has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Nano | Llama 4 Scout | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 33.5 | 27.7 |
| Released | 2025-08-07 | 2025-04-05 |
| Weights | Proprietary | Open |
| Context window | 400K | 128K |
| Max output | 128K | 4K |
| Input $ / M tokens | $0.05 | $0.10 |
| Output $ / M tokens | $0.40 | $0.30 |
| Results tracked | 49 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 Nano leads
GPT-5 Nano: 33.6 (#254), Llama 4 Scout: 20.2 (#339)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| SWE-bench Verified (bash only) | 34.8% | 9.1% |
| LMArena Coding | 1351 | 1286 |
| SciCode | — | 17% |
| WeirdML | 38.1% | — |
| BigCodeBench Complete | — | 43.1% |
| ALE-Bench | 718.67 | — |
Agentic & Tool Use GPT-5 Nano leads
GPT-5 Nano: 25.8 (#106), Llama 4 Scout: 24.6 (#119)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| Berkeley Function Calling Leaderboard | 51.5% | 28.1% |
| Terminal-Bench | 21.8% | — |
Reasoning GPT-5 Nano leads
GPT-5 Nano: 16.3 (#306), Llama 4 Scout: 9.1 (#345)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| ARC-AGI-2 | 2.6% | 0% |
| Kagi LLM Benchmark | 62.2% | 36.9% |
| ARC-AGI-1 | 20.7% | 0.5% |
| LMArena Hard Prompts | 1328 | 1266 |
| DTBench | 62.7% | 57.9% |
| LMCA | 7.9% | 12% |
| Epoch Capabilities Index | 139.38 | 129.64 |
| ForecastBench | 59.1 | 57.5 |
| CritPt | — | 0% |
| Chess Puzzles | 27% | — |
| Mystery Game Puzzles | 9% | — |
Math GPT-5 Nano leads
GPT-5 Nano: 29.4 (#241), Llama 4 Scout: 19.6 (#286)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 81.1% | 7.8% |
| Omni-MATH | 54.6% | 37.3% |
| LMArena Math | 1317 | 1287 |
| MATH Level 5 | 95.2% | 62.3% |
| FrontierMath (Feb 2025 set) | 8.3% | 0% |
| FrontierMath (Tiers 1-3) | 20% | — |
| FrontierMath Tier 4 | 2.4% | — |
| ProofBench | 12% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-5 Nano leads
GPT-5 Nano: 35.9 (#178), Llama 4 Scout: 31.9 (#217)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| GPQA Diamond | 69.4% | 51.8% |
| MMLU-Pro | 77.8% | 74.2% |
| Vectara Hallucination Rate | 10.5% | 7.7% |
| GPQA (HELM) | 67.9% | 50.7% |
| LMArena Expert | 1321 | 1235 |
| SimpleQA Verified | 11.7% | — |
Multimodal Too close to call
GPT-5 Nano: 31.3 (#108), Llama 4 Scout: 32.2 (#102)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| LMArena Vision | 1159 | 1118 |
| VPCT | 37.2% | — |
| SpatialViz-Bench | — | 34.2% |
Multilingual GPT-5 Nano leads
GPT-5 Nano: 45.3 (#172), Llama 4 Scout: 41.0 (#212)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| LMArena Non-English | 1313 | 1252 |
| LMArena Chinese | 1356 | 1255 |
| LMArena German | 1327 | 1272 |
| LMArena Japanese | 1226 | 1206 |
| LMArena Korean | 1269 | 1207 |
| LMArena Russian | 1296 | 1263 |
| LMArena Spanish | 1360 | 1278 |
| LMArena French | — | 1282 |
Instruction Following GPT-5 Nano leads
GPT-5 Nano: 75.0 (#79), Llama 4 Scout: 65.8 (#217)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| IFEval | 93.2% | 81.8% |
| LMArena Instruction Following | 1306 | 1248 |
Long Context GPT-5 Nano leads
GPT-5 Nano: 31.3 (#281), Llama 4 Scout: 27.5 (#294)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| Fiction.LiveBench | 44.4% | 36% |
| LMArena Longer Query | 1312 | 1265 |
Writing & Preference GPT-5 Nano leads
GPT-5 Nano: 39.1 (#249), Llama 4 Scout: 37.0 (#261)
| Benchmark | GPT-5 Nano | Llama 4 Scout |
|---|---|---|
| LMArena Text | 1320 | 1279 |
| LMArena Creative Writing | 1249 | 1249 |
| EQ-Bench Creative Writing | 705 | 783 |
| WildBench | 80.6% | 78% |
| LMArena Multi-Turn | 1311 | 1280 |
Frequently asked questions
Is GPT-5 Nano better than Llama 4 Scout?
GPT-5 Nano is the stronger model overall, scoring 33.5 to 27.7 on the Noometry Index.
Which is cheaper, GPT-5 Nano or Llama 4 Scout?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Llama 4 Scout lists at $0.10 and $0.30.
Is GPT-5 Nano or Llama 4 Scout better for coding?
GPT-5 Nano scores higher on coding benchmarks: 33.6 versus 20.2 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 128K.
How many benchmarks do GPT-5 Nano and Llama 4 Scout share?
38 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and Llama 4 Scout has 43.