Across 4 comparable benchmarks, GLM-4.5-Air has better reported results on 4 and worse on 0.
Moderate evidence4 of 4 shared results countedHow we compare
Price basis: Listed rates vs international rates. The price section below shows the applicable periods.
| Fact | QwQ-32B | GLM-4.5-Air |
|---|---|---|
| Context window | 128K tokens | 128K tokens |
| Input / 1M | — | $0.200 |
| Output / 1M | — | $1.10 |
| Blended (4:1) | — | $0.258 |
| Status | Available | Available |
| Released | March 6, 2025 | July 28, 2025 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Mean of 3 shared benchmarks grouped by capability on a 0–100 scale. How averages are built
GLM-4.5-Air ahead on 2 of 2 counted
GLM-4.5-Air ahead on 1 of 1 counted
No shared comparable results for Coding, Abstract reasoning, Computer & tool use, Research & search, Long context, Multimodal, Professional work & productivity, Cybersecurity.
Published scores, with matching modes where available. Gaps show QwQ-32B relative to GLM-4.5-Air. Reading the results
| Benchmark | QwQ-32B | GLM-4.5-Air | Gap |
|---|---|---|---|
| AIME 2024Mathematics · Accuracy | 79.5Normal · no tools | 89.4≠Thinking · no tools | −9.9 pts#25 vs #12 of 54 |
| MATH-500Mathematics · Accuracy | 91.0Normal · no tools | 98.1≠Thinking · no tools | −7.1 pts#41 vs #15 of 49 |
| GPQA DiamondScientific Reasoning · Accuracy | 58.0Normal · no tools | 75.0≠Thinking · no tools | −17.0 pts#168 vs #123 of 200 |
| MMLU-ProKnowledge Exams · Accuracy | 76.0Normal · no tools | 81.4≠Thinking · no tools | −5.4 pts#79 vs #53 of 153 |
Scientific Reasoning · Accuracy
Knowledge Exams · Accuracy
A dash means no result is available.
| Benchmark | QwQ-32B | GLM-4.5-Air |
|---|---|---|
| HLE | — | 10.6 |
| Aider-Polyglot | 20.9 | — |
| HumanEval | 19.0 | — |
| LiveCodeBench | — | 70.7 |
| SWE-bench Verified | — | 57.6 |
| Terminal Bench Hard | — | 20.5 |
| Terminal-Bench 1.0 | — | 30.0 |
| Pinch Bench | — | 85.7 |
| ECI | 138 | — |
| Creative Writing | 1258 | — |
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix · 5% steps
Blended price = effective input rate at 90% cache hits × 80% + output rate × 20%
A blended cost needs a published input and output price on both sides, and one of these two does not have them yet.
QwQ-32B: Listed text rates; pricing tier not confirmed.
GLM-4.5-Air: International pricing. Z.ai official (OpenRouter Z.AI first-party row), verified 2026-09-25
| Price / 1M tokens | QwQ-32B | GLM-4.5-Air | Ratio |
|---|---|---|---|
| Input | — | $0.20 | — |
| Output | — | $1.10 | — |
| Cached input | — | $0.03 | — |
| Cache write | — | — | — |
| Blended | — | $0.258 | — |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
| Attribute | QwQ-32B | GLM-4.5-Air |
|---|---|---|
| Availability | Available | Available |
| Provider | Alibaba | Zhipu AI |
| Model type | Reasoning model | Reasoning model |
| First release | March 6, 2025 | July 28, 2025 |
| Context window | 128K tokens | 128K tokens |
| Max output | — | 95K tokens |
| Total parameters | 32.5B | 106B |
| Active parameters | — | 12B |
| Architecture | Dense | Mixture of Experts (MoE) |
| Documented capabilities | Reasoning, Multilingual | Reasoning, Multilingual |
| Input modalities | Text | Text |
| Output modalities | Text | Text |
| Weights | Open weights | Open weights |
| Licence | Apache 2.0 | MIT License |
| Commercial use | Commercial use allowed | Commercial use allowed |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used QwQ-32B or GLM-4.5-Air? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.