Model comparison
GPT-5.4 vs Olmo 3.1 32b Instruct
GPT-5.4 is the stronger model overall, scoring 59.4 to 39.4 on the Noometry Index.
Last verified . 16 shared benchmarks.
Allen Institute for AI (Ai2)Summary
- They share 16 benchmarks with published results for both. GPT-5.4 scores higher in 8 categories and Olmo 3.1 32b Instruct in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 leads 73.5 to 36.3.
- Olmo 3.1 32b Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 | Olmo 3.1 32b Instruct | |
|---|---|---|
| Provider | OpenAI | Allen Institute for AI (Ai2) |
| Noometry Index | 59.4 | 39.4 |
| Released | 2026-03-05 | — |
| Weights | Proprietary | Open |
| Context window | 1.05M | — |
| Max output | 128K | — |
| Input $ / M tokens | $2.50 | — |
| Output $ / M tokens | $15 | — |
| Results tracked | 68 | 16 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 leads
GPT-5.4: 52.6 (#33), Olmo 3.1 32b Instruct: 39.5 (#157)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Coding | 1497 | 1347 |
| SWE-bench Verified | 76.9% | — |
| DeepSWE | 51.8% | — |
| LMArena WebDev | 1465 | — |
| SciCode | 56.6% | — |
| GSO | 31.4% | — |
| WeirdML | 77.7% | — |
| MirrorCode | 15.6% | — |
| ALE-Bench | 1,607 | — |
| AlgoTune | 1.85 | — |
Agentic & Tool Use Not comparable
GPT-5.4: 46.5 (#13), Olmo 3.1 32b Instruct: —
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| Terminal-Bench | 81.8% | — |
| APEX-Agents | 52.4% | — |
| τ²-bench Banking | 39.4% | — |
| DeepResearch Bench | 35.1% | — |
| PostTrainBench | 19% | — |
| GBAEval | 45.1% | — |
| LMArena Search | 1197 | — |
| METR Time Horizons | 74.3% | — |
| Vending-Bench 2 | 6,144 | — |
Reasoning GPT-5.4 leads
GPT-5.4: 61.8 (#19), Olmo 3.1 32b Instruct: 26.4 (#132)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Hard Prompts | 1485 | 1322 |
| ARC-AGI-2 | 74% | — |
| Kagi LLM Benchmark | 63.8% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 93.7% | — |
| CritPt | 23.4% | — |
| Chess Puzzles | 44% | — |
| EnigmaEval | 16% | — |
| Thematic Generalization | 80% | — |
| EBR-Bench | 25.4% | — |
| Mystery Game Puzzles | 37% | — |
| DTBench | 94.4% | — |
| LMCA | 52% | — |
| Epoch Capabilities Index | 156.81 | — |
| ForecastBench | 59.5 | — |
Math GPT-5.4 leads
GPT-5.4: 73.5 (#19), Olmo 3.1 32b Instruct: 36.3 (#167)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Math | 1488 | 1305 |
| FrontierMath (Tiers 1-3) | 78.6% | — |
| FrontierMath Tier 4 | 49% | — |
| MathArena Final-Answer Competitions | 83.1% | — |
| OTIS Mock AIME 2024-2025 | 97.8% | — |
| ProofBench | 56% | — |
| FrontierMath (Feb 2025 set) | 47.6% | — |
| FrontierMath Tier 4 (v1) | 27.1% | — |
Knowledge GPT-5.4 leads
GPT-5.4: 65.3 (#14), Olmo 3.1 32b Instruct: 36.1 (#175)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Expert | 1507 | 1308 |
| GPQA Diamond | 93.3% | — |
| Humanity's Last Exam | 36.2% | — |
| SimpleQA Verified | 45.1% | — |
| Vectara Hallucination Rate | 7% | — |
Multimodal Not comparable
GPT-5.4: 43.7 (#20), Olmo 3.1 32b Instruct: —
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Vision | 1303 | — |
| Blueprint-Bench 2 | 27.1% | — |
| Furniture Assembly | 37.5% | — |
| LMArena Document | 1471 | — |
Multilingual GPT-5.4 leads
GPT-5.4: 56.2 (#23), Olmo 3.1 32b Instruct: 42.6 (#191)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Non-English | 1465 | 1275 |
| LMArena Chinese | 1519 | 1304 |
| LMArena French | 1493 | 1328 |
| LMArena German | 1472 | 1282 |
| LMArena Korean | 1448 | 1206 |
| LMArena Russian | 1480 | 1268 |
| LMArena Spanish | 1454 | 1336 |
| LMArena Japanese | 1485 | — |
Instruction Following GPT-5.4 leads
GPT-5.4: 77.1 (#27), Olmo 3.1 32b Instruct: 68.6 (#187)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Instruction Following | 1469 | 1299 |
Long Context GPT-5.4 leads
GPT-5.4: 50.3 (#8), Olmo 3.1 32b Instruct: 39.9 (#166)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Longer Query | 1473 | 1312 |
| CL-bench | 27.9% | — |
| CL-bench Life | 21.7% | — |
Writing & Preference GPT-5.4 leads
GPT-5.4: 71.9 (#17), Olmo 3.1 32b Instruct: 50.2 (#185)
| Benchmark | GPT-5.4 | Olmo 3.1 32b Instruct |
|---|---|---|
| LMArena Text | 1469 | 1311 |
| LMArena Creative Writing | 1439 | 1264 |
| LMArena Multi-Turn | 1482 | 1309 |
| EQ-Bench Creative Writing | 1840 | — |
| EQ-Bench 4 | 1272 | — |
Frequently asked questions
Is GPT-5.4 better than Olmo 3.1 32b Instruct?
GPT-5.4 is the stronger model overall, scoring 59.4 to 39.4 on the Noometry Index.
Is GPT-5.4 or Olmo 3.1 32b Instruct better for coding?
GPT-5.4 scores higher on coding benchmarks: 52.6 versus 39.5 in the Noometry coding category.
How many benchmarks do GPT-5.4 and Olmo 3.1 32b Instruct share?
16 benchmarks have published results for both models. GPT-5.4 has 68 scored results on Noometry and Olmo 3.1 32b Instruct has 16.