Across 8 comparable benchmarks, GPT-5.1 Codex has better reported results on 7 and worse on 1.
Moderate evidence8 of 8 shared results countedHow we compare
GPT-5.1 Codex is no longer available through its official API. API prices for this version are not shown.
Price basis: International rates vs listed rates. The price section below shows the applicable periods.
| Fact | GLM-4.6 | GPT-5.1 Codex |
|---|---|---|
| Context window | 200K tokens | 400K tokens |
| Input / 1M | $0.600 | — |
| Output / 1M | $2.20 | — |
| Blended (4:1) | $0.567 | — |
| Status | Available | Retired |
| Released | September 30, 2025 | November 14, 2025 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Mean of 5 shared benchmarks grouped by capability on a 0–100 scale. How averages are built
GPT-5.1 Codex ahead on 5 of 5 counted
GLM-4.6 ahead on 1 of 1 counted
GPT-5.1 Codex ahead on 1 of 1 counted
No fixed-scale average — see the rows in Benchmark results
No shared comparable results for Abstract reasoning, Computer & tool use, Research & search, Long context, Multimodal, Professional work & productivity, Cybersecurity.
Published scores, with matching modes where available. Gaps show GLM-4.6 relative to GPT-5.1 Codex. Reading the results
| Benchmark | GLM-4.6 | GPT-5.1 Codex | Gap |
|---|---|---|---|
| LiveCodeBenchAlgorithmic Coding · Pass @K | 84.5Thinking · tools | 85.5≠Thinking High · tools | 0.99×#27 vs #24 of 113 |
| SWE-bench VerifiedRepository Engineering · Accuracy | 68.0Normal · no tools | 70.4≠Thinking High · tools | −2.4 pts#68 vs #60 of 108 |
| Terminal Bench HardAgentic Development · Accuracy | 28.8Normal · tools | 34.8≠Thinking High · tools | −6.0 pts#72 vs #50 of 157 |
| Terminal-Bench 1.0Agentic Development · Accuracy | 40.5Thinking · tools | 56.3≠Thinking High · tools | −15.8 pts#10 vs #1 of 26 |
| Vibe Code Bench v1.1Code Generation & Editing · Pass rate (%) | 3.1Unspecified Reasoning · tools | 13.1≠Thinking High · tools | −10.0 pts#113 vs #99 of 121 |
| AIME2025Mathematics · Accuracy | 98.6Thinking · no tools | 95.7≠Thinking High · no tools | +2.9 pts#15 vs #22 of 86 |
| CritPtScientific Reasoning · Score | 1.1Thinking · no tools | 5.7≠Thinking High · no tools | 0.19×#103 vs #66 of 134 |
| LiveBenchCross-capability Suites · Accuracy | 55.2Normal · no tools | 68.6Normal · no tools | −13.4 pts#91 vs #63 of 106 |
Algorithmic Coding · Pass @K
Repository Engineering · Accuracy
Agentic Development · Accuracy
Agentic Development · Accuracy
Code Generation & Editing · Pass rate (%)
Mathematics · Accuracy
Scientific Reasoning · Score
Cross-capability Suites · Accuracy
A dash means no result is available.
| Benchmark | GLM-4.6 | GPT-5.1 Codex |
|---|---|---|
| FrontierMath - Tier 4 | 2.1 | — |
| GPQA Diamond | 82.9 | — |
| HLE | 30.4 | — |
| MMLU-Pro | 83.2 | — |
| APEX-SWE | — | 29.0 |
| Terminal Bench 2.0 | — | 52.8 |
| Creative Writing | 1412 | — |
| MMMU-Pro | — | 72.5 |
| τ²-Bench | 75.9 | — |
| τ²-Bench - Telecom | 70.5 | — |
| τ³-Banking | 13.4 | — |
| BrowseComp | 45.1 | — |
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.
GLM-4.6: International pricing. Z.ai official (OpenRouter Z.AI first-party row), verified 2026-09-25
GPT-5.1 Codex: Listed text rates; pricing tier not confirmed.
| Price / 1M tokens | GLM-4.6 | GPT-5.1 Codex | Ratio |
|---|---|---|---|
| Input | $0.60 | — | — |
| Output | $2.20 | — | — |
| Cached input | $0.11 | — | — |
| Cache write | — | — | — |
| Blended | $0.567 | — | — |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
Reported cost per task, paired with the score from the same run. Settings and sources may differ; costs do not affect the verdict.
| Benchmark | GLM-4.6 | GPT-5.1 Codex |
|---|---|---|
| Vibe Code Bench v1.1 | Not reported |
| Attribute | GLM-4.6 | GPT-5.1 Codex |
|---|---|---|
| Availability | Available | Retired |
| Provider | Zhipu AI | OpenAI |
| Model type | Chat model | Coding model |
| First release | September 30, 2025 | November 14, 2025 |
| Context window | 200K tokens | 400K tokens |
| Max output | 128K tokens | 128K tokens |
| Total parameters | 355B | — |
| Active parameters | 32B | — |
| Architecture | Mixture of Experts (MoE) | Undisclosed |
| Documented capabilities | Reasoning, Multilingual | Reasoning, Coding |
| Input modalities | Text | Text, Image |
| Output modalities | Text | Text |
| Weights | Open weights | Closed / API only |
| Licence | MIT License | — |
| Commercial use | Commercial use allowed | Proprietary |
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 GLM-4.6 or GPT-5.1 Codex? 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.