GLM-4.6: benchmark scores, price and ranking
GLM-4.6 from Z.ai (Zhipu AI),Tsinghua University ranks #610 of 743 LLMs on the LLMs Tiger aggregate benchmark score (45.0), based on 6 benchmark results captured up to 2026-10-13.
- Organisation
- Z.ai (Zhipu AI),Tsinghua University
- Type
- Open-weight / local model
- Open weights
- Yes
- Released
- 2025-09-30 (source: curated)
- Cheapest API price
- $0.450 in / $1.80 out per million tokens (via vercel_ai_gateway, observed 2026-10-13T20:47Z)
- Aggregate rank
- #610 of 743 (score 45.0)
Relative to other models, GLM-4.6 is strongest on SWE-bench Verified (#38 of 73) and weakest on ALE-Bench (#68 of 79).
Across the 6 benchmarks where at least ten models are scored, GLM-4.6 is in the top 10% on 0 and in the top half on 0.
Its cheapest listed input price, $0.450 per million tokens, is lower than that of 52% of the 427 models with a listed price.
GLM-4.6 benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| SWE-bench Verified | 55.4% | #38 of 73 | SWE-bench Verified leaderboard | 2026-10-13 |
| CritPt | 1.1% | #66 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 38.4% | #72 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WebDev Arena | 1,341 | #47 of 67 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Terminal-Bench 2.0 | 24.5% | #24 of 28 | Epoch AI Benchmarking Hub | 2026-10-13 |
| ALE-Bench | 341 | #68 of 79 | Epoch AI Benchmarking Hub | 2026-10-13 |
About GLM-4.6
GLM-4.6 is an open-weight model that can run on your own hardware from Z.ai (Zhipu AI),Tsinghua University. It was released on 2025-09-30, according to the developer's announcement. The cheapest API listing we track charges $0.450 per million input tokens and $1.80 per million output tokens, through vercel_ai_gateway. We hold 6 benchmark results for it across 2 areas, and none is in the top tenth of its board.
Where GLM-4.6 stands
On our aggregate score, GLM-4.6 is #610 of 743 ranked models, in the lower half. Its score is 45.0; the scale is centred on 50, so it sits 5.0 points below the typical model. That beats 18% of the ranked field. Among the 4 ranked Z.ai (Zhipu AI),Tsinghua University models it is #4. The next one up is glm-4.5 at 48.4.
GLM-4.6, benchmark by benchmark
GLM-4.6 lands below the middle on SWE-bench Verified: 55.4%, #38 of 73, with 51% of the field scoring higher. Claude Opus 4.7 (max) leads at 83.5%; the closest result above is Devstral Small (2512) at 56.4%. It covers real GitHub issues resolved. The board median is 56.4%, so GLM-4.6 is 1.0 points below it. Source: SWE-bench Verified leaderboard, captured 2026-10-13.
GLM-4.6 lands below the middle on CritPt: 1.1%, #66 of 115, with 57% of the field scoring higher. GPT-5.6 Sol (max) leads at 32.3%; the closest result above is o3 (high) at 1.4%. It covers unpublished research-level physics problems. The board median is 2.9%, so GLM-4.6 is 1.7 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
GLM-4.6 lands below the middle on SciCode: 38.4%, #72 of 106, with 67% of the field scoring higher. Claude Opus 5.5 (max) leads at 66.9%; the closest result above is Mercury 2 at 38.7%. It covers scientific research code. The board median is 44.3%, so GLM-4.6 is 5.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
GLM-4.6 lands below the middle on WebDev Arena: 1,341, #47 of 67, with 69% of the field scoring higher. Claude Opus 5.5 (max) leads at 1,820; the closest result above is Laguna M.1 at 1,347. It covers building web apps, crowd-voted (Elo). The board median is 1,433, so GLM-4.6 is 93 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Terminal-Bench 2.0 is a weak spot. GLM-4.6 scores 24.5%, #24 of 28, behind 82% of the models measured. The leader, Gemini 3 Pro Preview, reaches 69.4%. It covers real terminal tasks (agent + model). The board median is 37.3%, so GLM-4.6 is 12.8 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
ALE-Bench is a weak spot. GLM-4.6 scores 341, #68 of 79, behind 85% of the models measured. The leader, GPT-6 Astra (max), reaches 2,951. It covers AtCoder heuristic-optimisation contests (performance rating). The board median is 764, so GLM-4.6 is 423 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Software engineering: weak, ahead of 28% of the field on average across 5 benchmarks. Its best board here is SWE-bench Verified (#38 of 73) and its weakest is ALE-Bench (#68 of 79).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so GLM-4.6 costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on WizardLM-2-8x22B ($0.480) and SWE-1.6 ($0.500).
Release history
The Z.ai (Zhipu AI),Tsinghua University model released before it was glm-4.5 on 2025-07-28, which scores 48.4.
Reading these numbers
The 6 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (5), SWE-bench Verified leaderboard (1). 1 of them come from an official leaderboard and 5 from datasets or third-party runs. With 6 results the aggregate rank can move a lot when one new score arrives.
Models ranked nearby
- #607 DeepSeek-V3 (45.1)
- #608 DeepSeek-Coder-V2-Lite-Base (45.0)
- #609 mistral-small-3-2501 (45.0)
- #611 qwen2.5-32b-instruct (45.0)
- #612 Qwen3-30B-A3B (44.9)
- #613 aya-expanse-8b (44.9)
More from Z.ai (Zhipu AI),Tsinghua University
- GLM-4.5 Thinking (#283, 49.2)
- GLM-4.5-Air (#327, 48.7)
- glm-4.5 (#351, 48.4)
Compare it on the live leaderboard → · How the aggregate score works