Model comparison
gpt-oss-120b vs Llama 3.1-8B
gpt-oss-120b is the stronger model overall, scoring 36.3 to 23.0 on the Noometry Index.
Last verified . 32 shared benchmarks.
MetaSummary
- They share 32 benchmarks with published results for both. gpt-oss-120b scores higher in 7 categories and Llama 3.1-8B in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where gpt-oss-120b leads 52.5 to 10.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.9% for gpt-oss-120b and 1.7% for Llama 3.1-8B.
- Llama 3.1-8B is cheaper at $0.05 / $0.08 per million input/output tokens, against $0.037 / $0.17 for gpt-oss-120b.
- gpt-oss-120b accepts more context: 131K tokens versus 128K.
Side by side
| gpt-oss-120b | Llama 3.1-8B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 36.3 | 23.0 |
| Released | 2025-08-05 | 2024-07-23 |
| Weights | Open | Open |
| Context window | 131K | 128K |
| Max output | 41K | 4K |
| Input $ / M tokens | $0.037 | $0.05 |
| Output $ / M tokens | $0.17 | $0.08 |
| Results tracked | 48 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding gpt-oss-120b leads
gpt-oss-120b: 33.5 (#256), Llama 3.1-8B: 20.2 (#340)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| SciCode | 36% | 13.2% |
| WeirdML | 48.2% | 1.7% |
| LMArena Coding | 1380 | 1195 |
| SWE-bench Verified (bash only) | 26% | — |
| Aider Polyglot | 41.8% | — |
| BigCodeBench Instruct | — | 32.8% |
| BigCodeBench Complete | — | 40.5% |
| ALE-Bench | 575.62 | — |
| AlgoTune | 1.41 | — |
| HumanEval+ | — | 62.8% |
| MBPP+ | — | 55.6% |
Agentic & Tool Use Llama 3.1-8B leads
gpt-oss-120b: 12.2 (#153), Llama 3.1-8B: 22.5 (#131)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| Terminal-Bench | 18.7% | — |
| APEX-Agents | 4.4% | — |
| Berkeley Function Calling Leaderboard | — | 25.8% |
| BALROG | — | 15.1% |
| METR Time Horizons | 56.6% | — |
| Vending-Bench 2 | -21.53 | — |
Reasoning gpt-oss-120b leads
gpt-oss-120b: 20.0 (#245), Llama 3.1-8B: 14.9 (#321)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| CritPt | 1.1% | 0% |
| Chess Puzzles | 20% | 0% |
| LMArena Hard Prompts | 1364 | 1175 |
| DTBench | 76.3% | 50.9% |
| LMCA | 22.1% | 5.4% |
| Epoch Capabilities Index | 139.93 | 116.57 |
| SimpleBench | 22.1% | — |
| Kagi LLM Benchmark | 58.6% | — |
| Mystery Game Puzzles | 2% | — |
| Surface Evolver Bench | 25% | — |
| PIQA | — | 81.2% |
Math gpt-oss-120b leads
gpt-oss-120b: 52.5 (#50), Llama 3.1-8B: 10.2 (#317)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 1.7% |
| Omni-MATH | 68.8% | 13.7% |
| LMArena Math | 1389 | 1179 |
| MATH Level 5 | — | 22.9% |
| GSM8K | — | 82.4% |
Knowledge gpt-oss-120b leads
gpt-oss-120b: 42.4 (#96), Llama 3.1-8B: 8.0 (#307)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| GPQA Diamond | 75.8% | 27% |
| MMLU-Pro | 79.5% | 40.6% |
| GPQA (HELM) | 68.4% | 24.7% |
| LMArena Expert | 1356 | 1144 |
| Confabulations | 15.7% | — |
| Vectara Hallucination Rate | 14.2% | — |
| BoolQ | — | 82.8% |
| MMLU | — | 56.1% |
Multilingual gpt-oss-120b leads
gpt-oss-120b: 48.0 (#147), Llama 3.1-8B: 34.0 (#249)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| LMArena Non-English | 1351 | 1148 |
| LMArena Chinese | 1385 | 1151 |
| LMArena French | 1369 | 1177 |
| LMArena German | 1353 | 1144 |
| LMArena Japanese | 1331 | 1061 |
| LMArena Korean | 1282 | 1053 |
| LMArena Russian | 1343 | 1158 |
| LMArena Spanish | 1389 | 1169 |
Instruction Following gpt-oss-120b leads
gpt-oss-120b: 69.3 (#173), Llama 3.1-8B: 58.9 (#258)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| IFEval | 83.6% | 74.3% |
| LMArena Instruction Following | 1318 | 1159 |
Long Context Llama 3.1-8B leads
gpt-oss-120b: 31.4 (#278), Llama 3.1-8B: 35.8 (#238)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| LMArena Longer Query | 1319 | 1182 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference gpt-oss-120b leads
gpt-oss-120b: 46.5 (#217), Llama 3.1-8B: 29.7 (#290)
| Benchmark | gpt-oss-120b | Llama 3.1-8B |
|---|---|---|
| LMArena Text | 1365 | 1187 |
| LMArena Creative Writing | 1275 | 1154 |
| EQ-Bench Creative Writing | 961 | 713 |
| WildBench | 84.5% | 68.7% |
| LMArena Multi-Turn | 1340 | 1172 |
| Short-Story Creative Writing | 77.1% | — |
Frequently asked questions
Is gpt-oss-120b better than Llama 3.1-8B?
gpt-oss-120b is the stronger model overall, scoring 36.3 to 23.0 on the Noometry Index.
Which is cheaper, gpt-oss-120b or Llama 3.1-8B?
Llama 3.1-8B is cheaper. It lists at $0.05 per million input tokens and $0.08 per million output tokens; gpt-oss-120b lists at $0.037 and $0.17.
Is gpt-oss-120b or Llama 3.1-8B better for coding?
gpt-oss-120b scores higher on coding benchmarks: 33.5 versus 20.2 in the Noometry coding category.
Which has the bigger context window?
gpt-oss-120b does, with 131K tokens against 128K.
How many benchmarks do gpt-oss-120b and Llama 3.1-8B share?
32 benchmarks have published results for both models. gpt-oss-120b has 48 scored results on Noometry and Llama 3.1-8B has 43.