Model comparison
GPT-5.1 vs Llama-3.3-70B-Instruct
GPT-5.1 is the stronger model overall, scoring 49.0 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 22× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.
Last verified . 35 shared benchmarks.
MetaSummary
- They share 35 benchmarks with published results for both. GPT-5.1 scores higher in 9 categories and Llama-3.3-70B-Instruct in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.1 leads 52.2 to 15.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.6% for GPT-5.1 and 5.1% for Llama-3.3-70B-Instruct.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $1.25 / $10 for GPT-5.1.
- GPT-5.1 accepts more context: 400K tokens versus 128K.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-5.1 | Llama-3.3-70B-Instruct | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 49.0 | 30.6 |
| Released | 2025-11-13 | 2024-12-06 |
| Weights | Proprietary | Open |
| Context window | 400K | 128K |
| Max output | 128K | 4K |
| Input $ / M tokens | $1.25 | $0.10 |
| Output $ / M tokens | $10 | $0.32 |
| Results tracked | 63 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.1 leads
GPT-5.1: 46.4 (#66), Llama-3.3-70B-Instruct: 31.0 (#290)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| SciCode | 43.3% | 26% |
| WeirdML | 60.8% | 14.4% |
| LiveBench Coding | 72.5% | 36.6% |
| LMArena Coding | 1454 | 1268 |
| SWE-bench Verified | 68% | — |
| SWE-bench Verified (bash only) | 66% | — |
| LMArena WebDev | 1395 | — |
| GSO | 13.7% | — |
| BigCodeBench Instruct | — | 46.9% |
| BigCodeBench Complete | — | 57.5% |
| ALE-Bench | 1,192 | — |
Agentic & Tool Use GPT-5.1 leads
GPT-5.1: 32.7 (#60), Llama-3.3-70B-Instruct: 25.8 (#105)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| Terminal-Bench | 47.6% | — |
| Berkeley Function Calling Leaderboard | — | 31.9% |
| DeepResearch Bench | 42.8% | — |
| BALROG | — | 23% |
| LMArena Search | 1199 | — |
| Vending-Bench 2 | 1,473 | — |
Reasoning GPT-5.1 leads
GPT-5.1: 39.8 (#58), Llama-3.3-70B-Instruct: 14.1 (#327)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| SimpleBench | 53.2% | 19.9% |
| CritPt | 4.9% | 0% |
| LiveBench Reasoning | 95.8% | 50.8% |
| LMArena Hard Prompts | 1457 | 1257 |
| DTBench | 90.1% | 59.5% |
| LiveBench Data Analysis | 72.1% | 49.5% |
| LMCA | 43.9% | 17.5% |
| Epoch Capabilities Index | 149.64 | 127.33 |
| ForecastBench | 58.1 | 58.6 |
| LiveBench | 78.8% | 50.2% |
| ARC-AGI-2 | 17.6% | — |
| ARC-AGI-1 | 72.8% | — |
| Chess Puzzles | 32% | — |
| EnigmaEval | 11.2% | — |
| Mystery Game Puzzles | 19% | — |
Math GPT-5.1 leads
GPT-5.1: 52.2 (#51), Llama-3.3-70B-Instruct: 15.3 (#298)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.6% | 5.1% |
| LiveBench Math | 94.5% | 42.2% |
| LMArena Math | 1447 | 1267 |
| Omni-MATH | 46.4% | — |
| MATH Level 5 | — | 41.6% |
| FrontierMath (Feb 2025 set) | 31% | — |
| FrontierMath Tier 4 (v1) | 12.5% | — |
Knowledge GPT-5.1 leads
GPT-5.1: 50.6 (#71), Llama-3.3-70B-Instruct: 30.6 (#226)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| GPQA Diamond | 87.6% | 47.4% |
| Vectara Hallucination Rate | 10.9% | 4.1% |
| LMArena Expert | 1470 | 1225 |
| Humanity's Last Exam | 23.7% | — |
| SimpleQA Verified | 48% | — |
| MMLU-Pro | 57.9% | — |
| Confabulations | — | 22.8% |
| GPQA (HELM) | 44.2% | — |
| MMLU | — | 86.3% |
Multimodal Not comparable
GPT-5.1: 44.8 (#19), Llama-3.3-70B-Instruct: —
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Vision | 1250 | — |
| VPCT | 58.7% | — |
| LMArena Document | 1403 | — |
Multilingual GPT-5.1 leads
GPT-5.1: 53.8 (#56), Llama-3.3-70B-Instruct: 39.9 (#220)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | 1431 | 1236 |
| LMArena Chinese | 1495 | 1217 |
| LMArena French | 1450 | 1281 |
| LMArena German | 1438 | 1251 |
| LMArena Japanese | 1453 | 1150 |
| LMArena Korean | 1401 | 1143 |
| LMArena Russian | 1435 | 1252 |
| LMArena Spanish | 1433 | 1270 |
Instruction Following GPT-5.1 leads
GPT-5.1: 83.9 (#1), Llama-3.3-70B-Instruct: 71.1 (#157)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Instruction Following | 93.3% | 82.7% |
| LMArena Instruction Following | 1443 | 1242 |
| IFEval | 93.5% | — |
Long Context GPT-5.1 leads
GPT-5.1: 47.6 (#14), Llama-3.3-70B-Instruct: 26.4 (#295)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Longer Query | 1447 | 1256 |
| Fiction.LiveBench | — | 33.3% |
| CL-bench | 23.7% | — |
| CL-bench Life | 17.3% | — |
Writing & Preference GPT-5.1 leads
GPT-5.1: 64.5 (#55), Llama-3.3-70B-Instruct: 47.6 (#207)
| Benchmark | GPT-5.1 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | 1443 | 1274 |
| LMArena Creative Writing | 1427 | 1250 |
| LMArena Multi-Turn | 1450 | 1280 |
| LiveBench Language | 80.2% | 39.2% |
| WildBench | 86.3% | — |
Frequently asked questions
Is GPT-5.1 better than Llama-3.3-70B-Instruct?
GPT-5.1 is the stronger model overall, scoring 49.0 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 22× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.
Which is cheaper, GPT-5.1 or Llama-3.3-70B-Instruct?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; GPT-5.1 lists at $1.25 and $10.
Is GPT-5.1 or Llama-3.3-70B-Instruct better for coding?
GPT-5.1 scores higher on coding benchmarks: 46.4 versus 31.0 in the Noometry coding category.
Which has the bigger context window?
GPT-5.1 does, with 400K tokens against 128K.
How many benchmarks do GPT-5.1 and Llama-3.3-70B-Instruct share?
35 benchmarks have published results for both models. GPT-5.1 has 63 scored results on Noometry and Llama-3.3-70B-Instruct has 43.