Compare / OpenAI vs Z.ai
OpenAI vs Z.ai: which AI lab is ahead?
Each lab's best model on sale goes head to head on every kind of work we measure, from our own benchmarks and licensed sources. 39 models in all; a lab is only as good as the model you would actually pick from it.
Kinds of work led by each lab's best model there, of the 10 both are measured on.
Results as of 19 November 2026 (catalogue release 2026-11-19-3608133e8331). Models on sale means listed on OpenRouter today.
OpenAI or Z.ai? The short answer
OpenAI is ahead: its best models beat Z.ai's in all 10 kinds of work both are measured on.
- Pick OpenAI for product listings, decks from an analysis, user surveys, marketing planning, agents and tool use, professional work, reasoning and knowledge, writing, sticking to the facts and speed under pressure.
- On our overall leaderboard, GPT-6.2 Astra is #2 (86 of 100) and GLM-5.3 is #30 (35).
- Prices start at $0.058 per million tokens for OpenAI (gpt-oss-20b) and $0.15 for Z.ai (GLM-4.7-Flash).
Head to head
Each line names the model behind the number. Prices are blended, three input tokens to one output, at the developer's list price where we have read it; see every model's price. Intelligence Index from Artificial Analysis; overall score from our leaderboard.
Which lab is better for what
For each kind of work, each lab's best model there and its rank among every model on sale. Further right is better; the better rank is in green.
Every model, by intelligence and price
Each dot is a model on sale with an Intelligence Index from Artificial Analysis. Up is smarter; left is cheaper. Hover or tap a dot for its name.
OpenAIZ.ai
On our overall leaderboard
The 10 of their models we rank overall: one score out of 100 across everything we publish (how the score works). Only models with enough of our own results are ranked.
- #2GPT-6.2 Astra86
- #3GPT-6.2 Sol84
- #5GPT-6.1 Sol78
- #6GPT-6.1 Terra77
- #7GPT-6 Astra76
- #10GPT-6 Sol66
- #27GPT-6 Luna38
- #30GLM-5.335
- #32GLM-5.3 FlashX Legal29
- #38GLM-5V-Turbo12
Benchmark by benchmark
The headline result of every benchmark both labs have models on: each lab's best model and its rank among the models on sale. Green marks the better rank.
| Measure | OpenAI | Z.ai |
|---|---|---|
| Agents and tool use | ||
| τ²-bench · Task success (pass^1) · tau2 v1.0.1, retail% of tasks, higher is better | 81.6%GPT-5.2 · 2 of 6 | 73.7%GLM-5 · 5 of 6 |
| τ²-bench · Task success (pass^1) · tau2 v1.0.1, airline% of tasks, higher is better | 83.0%GPT-5.2 · 2 of 6 | 82.5%GLM-5 · 3 of 6 |
| τ²-bench · Task success (pass^1) · tau2 v1.0.1, telecom% of tasks, higher is better | 89.7%GPT-5.2 · 4 of 6 | 86.8%GLM-5 · 5 of 6 |
| Berkeley Function Calling Leaderboard (BFCL) V4 · Overall accuracy · BFCL% correct, higher is better | 63.1%o3 · 5 of 36 | 72.4%GLM-4.6 · 3 of 36 |
| OpenHands Index · Average% resolved, higher is better | 65.9%GPT-5.5 · 5 of 31 | 58.2%GLM-5.1 · 12 of 31 |
| Artificial Analysis · Terminal-Bench 4.0% of tasks, higher is better | 59.6%GPT-6 Astra · 3 of 107 | 41.9%GLM-5.3 · 13 of 107 |
| Artificial Analysis · Terminal-Bench 2.1% of tasks, higher is better | 89.9%GPT-6 Astra · 2 of 107 | 84.3%GLM-5.3-Flash · 14 of 107 |
| Artificial Analysis · τ-bench banking% of tasks, higher is better | 44.3%GPT-5.6 Sol · 12 of 104 | 50.3%GLM-5.3 · 3 of 104 |
| Vending-Bench 2 · Money after a yearUS dollars, higher is better | $15,515GPT-6 Astra · 1 of 53 | $8,314GLM-5.2 · 7 of 53 |
| Decks from an analysis | ||
| DeckBench · Deck ratingrating, higher is better | 1,551GPT-6 Astra · 2 of 35 | 809GLM-5.3 FlashX Legal · 27 of 35 |
| Marketing planning | ||
| ROASBench · Overall scorescore out of 100, higher is better | 58.4GPT-6.1 Terra · 2 of 35 | 46.2GLM-5.3 FlashX Legal · 17 of 35 |
| Product listings | ||
| CatalogBench · Reliably publish-ready% of products, higher is better | 75.0%GPT-6.2 Sol · 2 of 35 | 1.8%GLM-5V-Turbo · 32 of 35 |
| CatalogBench · Reliably publish-ready · sales brief% of products, higher is better | 75.0%GPT-6.2 Sol · 1 of 35 | 0.0%GLM-5.3 FlashX Legal · 30 of 35 |
| Professional work | ||
| APEX-Agents · Tasks passed% of tasks, higher is better | 64.7%GPT-6 Astra · 7 of 37 | 56.6%GLM-5.3 · 13 of 37 |
| GDP.pdf · Professional document tasks% of rubric, higher is better | 30.7%GPT-5.6 Sol · 1 of 31 | 14.0%GLM-5.3-Flash · 24 of 31 |
| Reasoning and knowledge | ||
| Artificial Analysis · Artificial Analysis Intelligence Indexindex score, higher is better | 52.7GPT-6 Astra · 4 of 185 | 44.8GLM-5.3 · 17 of 185 |
| Artificial Analysis · Humanity's Last Exam% of questions, higher is better | 54.7%GPT-6 Astra · 7 of 186 | 42.3%GLM-5.3 · 37 of 186 |
| Artificial Analysis · GPQA Diamond% of questions, higher is better | 96.3%GPT-6 Astra · 1 of 168 | 91.7%GLM-5.3 · 24 of 168 |
| Speed under pressure | ||
| BulletBench · Ladder Elo · Lightning 10+1ladder Elo, higher is better | 584GPT-6.1 Sol · 4 of 18 | 310GLM-5.3 FlashX Legal · 12 of 18 |
| BulletBench · Ladder Elo · Bullet 60sladder Elo, higher is better | 928GPT-6.1 Sol · 2 of 36 | 358GLM-5.3 FlashX Legal · 20 of 36 |
| Sticking to the facts | ||
| Vectara Hallucination Leaderboard · Hallucination rate · Vectara% of summaries, lower is better | 3.1%GPT-5.4 nano · 1 of 71 | 9.3%GLM-4.7-Flash · 34 of 71 |
| Arena (formerly LMArena) · Overall · Text factualityArena rating, higher is better | 1,486GPT-5.5 · 6 of 109 | 1,467GLM-5.3 · 25 of 109 |
| SimpleQA Verified (Epoch AI) · Correct answers% of questions, higher is better | 75.6%GPT-6 Astra · 1 of 68 | 41.0%GLM-5.3 · 38 of 68 |
| User surveys | ||
| SurveyBench · Analysis ratingrating, higher is better | 1,476GPT-6.2 Astra · 2 of 26 | 937GLM-5.3 FlashX Legal · 16 of 26 |
| Writing | ||
| Arena (formerly LMArena) · Overall · TextArena rating, higher is better | 1,484GPT-5.6 Sol · 15 of 153 | 1,478GLM-5.3 · 22 of 153 |
| UGI Leaderboard · Writing score · UGIscore out of 100, higher is better | 69.7GPT-5.5 · 11 of 124 | 67.0GLM-5.2 · 18 of 124 |
Quick answers
Is OpenAI or Z.ai better?
OpenAI is ahead: its best models beat Z.ai's in all 10 kinds of work both are measured on.
What is OpenAI's best model?
GPT-6.2 Astra: #2 on the Spring Prompt overall leaderboard.
What is Z.ai's best model?
GLM-5.3: #30 on the Spring Prompt overall leaderboard and 44.8 on the Artificial Analysis Intelligence Index.
Which is better for agents and tool use, OpenAI or Z.ai?
OpenAI: its best model there, GPT-6 Astra, ranks 4 of 159 models on sale, against 15 for Z.ai's GLM-5.3.
Which is better for reasoning and knowledge, OpenAI or Z.ai?
OpenAI: its best model there, GPT-6 Astra, ranks 3 of 187 models on sale, against 31 for Z.ai's GLM-5.3.
Which is better for writing, OpenAI or Z.ai?
OpenAI: its best model there, GPT-5.5, ranks 15 of 169 models on sale, against 24 for Z.ai's GLM-5.3.
Which is cheaper, OpenAI or Z.ai?
OpenAI's cheapest model with results, gpt-oss-20b, costs $0.058 per million tokens (three input to one output), against $0.15 for GLM-4.7-Flash from Z.ai.
Other labs compared
- Anthropic vs OpenAIBest model against best model
- Google vs OpenAIBest model against best model
- OpenAI vs xAIBest model against best model
- Meta vs OpenAIBest model against best model
- DeepSeek vs OpenAIBest model against best model
- Alibaba vs OpenAIBest model against best model
- Mistral AI vs OpenAIBest model against best model
- Anthropic vs Z.aiBest model against best model
Which model is best for your product?
Benchmarks aren't your task. Run models from OpenAI, Z.ai and the rest on your own prompt, and see the cheapest one that passes.