Model comparison
GLM-5 vs o3
o3 is the stronger model overall, scoring 47.5 to 46.1 on the Noometry Index. GLM-5 costs 2.3× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 34 shared benchmarks.
Summary
- They share 34 benchmarks with published results for both. GLM-5 scores higher in 4 categories and o3 in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where o3 leads 53.3 to 44.7.
- The biggest single-benchmark swing is Chess Puzzles: 10% for GLM-5 and 38% for o3.
- GLM-5 is cheaper at $1 / $3.20 per million input/output tokens, against $2 / $8 for o3.
- GLM-5 accepts more context: 205K tokens versus 200K.
- GLM-5 has downloadable open weights; the other is API-only.
Side by side
| GLM-5 | o3 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 46.1 | 47.5 |
| Released | 2026-02-11 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 205K | 200K |
| Max output | 131K | 100K |
| Input $ / M tokens | $1 | $2 |
| Output $ / M tokens | $3.20 | $8 |
| Results tracked | 45 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5 leads
GLM-5: 49.0 (#52), o3: 46.8 (#64)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| SWE-bench Verified | 72.1% | 62.3% |
| SWE-bench Verified (bash only) | 72.8% | 58.4% |
| WeirdML | 48.2% | 52.4% |
| LMArena Coding | 1461 | 1408 |
| ALE-Bench | 765.62 | 933.55 |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1434 | — |
| SWE-bench Multilingual | 69.7% | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |
Agentic & Tool Use o3 leads
GLM-5: 31.1 (#71), o3: 34.5 (#44)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| Terminal-Bench | 52.4% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Banking | 9.8% | — |
| τ²-bench Retail | 73.7% | — |
| τ²-bench Telecom | 86.8% | — |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 4,432 | — |
Reasoning o3 leads
GLM-5: 27.6 (#116), o3: 32.0 (#78)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| ARC-AGI-2 | 4.9% | 6.5% |
| SimpleBench | 53.2% | 53.1% |
| Kagi LLM Benchmark | 75% | 67.6% |
| ARC-AGI-1 | 44.7% | 60.8% |
| Chess Puzzles | 10% | 38% |
| LMArena Hard Prompts | 1452 | 1402 |
| Epoch Capabilities Index | 145.83 | 146.86 |
| ForecastBench | 61 | 62.5 |
| NYT Connections (extended) | 74.8% | — |
| CritPt | — | 1.4% |
| EnigmaEval | — | 13.1% |
| Mystery Game Puzzles | — | 29% |
| DTBench | — | 84.8% |
| LMCA | — | 39.7% |
Math o3 leads
GLM-5: 46.4 (#71), o3: 50.2 (#58)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 80% | 84.4% |
| LMArena Math | 1440 | 1426 |
| FrontierMath (Feb 2025 set) | 16.4% | 18.7% |
| FrontierMath Tier 4 (v1) | 2.1% | 2.1% |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| MathArena Final-Answer Competitions | 65.7% | — |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
Knowledge o3 leads
GLM-5: 52.3 (#64), o3: 54.6 (#52)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| GPQA Diamond | 87.8% | 81.8% |
| LMArena Expert | 1454 | 1402 |
| Humanity's Last Exam | — | 20.3% |
| SimpleQA Verified | — | 49.4% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| Vectara Hallucination Rate | 10.1% | — |
| GPQA (HELM) | — | 75.3% |
Multimodal Not comparable
GLM-5: —, o3: 41.4 (#36)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| LMArena Vision | — | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
Multilingual GLM-5 leads
GLM-5: 53.7 (#58), o3: 51.7 (#105)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| LMArena Non-English | 1430 | 1401 |
| LMArena Chinese | 1511 | 1437 |
| LMArena French | 1455 | 1430 |
| LMArena German | 1445 | 1420 |
| LMArena Japanese | 1416 | 1403 |
| LMArena Korean | 1423 | 1370 |
| LMArena Russian | 1436 | 1406 |
| LMArena Spanish | 1454 | 1395 |
Instruction Following GLM-5 leads
GLM-5: 75.2 (#67), o3: 72.8 (#127)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| LMArena Instruction Following | 1428 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
GLM-5: 44.7 (#60), o3: 53.3 (#6)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| CL-bench | 18.7% | 17.8% |
| LMArena Longer Query | 1446 | 1372 |
| Fiction.LiveBench | — | 88.9% |
Writing & Preference GLM-5 leads
GLM-5: 66.0 (#38), o3: 63.5 (#64)
| Benchmark | GLM-5 | o3 |
|---|---|---|
| LMArena Text | 1446 | 1410 |
| LMArena Creative Writing | 1439 | 1359 |
| EQ-Bench Creative Writing | 1601 | 1676 |
| LMArena Multi-Turn | 1456 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |
Frequently asked questions
Is GLM-5 better than o3?
o3 is the stronger model overall, scoring 47.5 to 46.1 on the Noometry Index. GLM-5 costs 2.3× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, GLM-5 or o3?
GLM-5 is cheaper. It lists at $1 per million input tokens and $3.20 per million output tokens; o3 lists at $2 and $8.
Is GLM-5 or o3 better for coding?
GLM-5 scores higher on coding benchmarks: 49.0 versus 46.8 in the Noometry coding category.
Which has the bigger context window?
GLM-5 does, with 205K tokens against 200K.
How many benchmarks do GLM-5 and o3 share?
34 benchmarks have published results for both models. GLM-5 has 45 scored results on Noometry and o3 has 63.