Agentic
Like-for-like- GLM-5.2
- 81.0
- Qwen3.6-35B-A3B
- 51.5
- Weighted basis
- 1 vs 1 rows
- Reading
- GLM-5.2 leads
Model comparison
GLM-5.2 vs Qwen3.6-35B-A3B
Updated July 31, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
21 confirmed releases in the last 30 daystrack changes
GLM-5.2 has the higher public score estimate, 62.94 versus 50.48, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
12 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Agentic work
Tool use, computer use, and multi-step task completion
GLM-5.2
GLM-5.2 leads on the same 1 weighted benchmark row.
Confidence: limited
Long documents
Prompts that approach the documented context limit
GLM-5.2
GLM-5.2 has the larger documented context window.
Confidence: documented
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
Chat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: rate-fallback
Repository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
6 GLM-5.2 only
12 Shared
29 Qwen3.6-35B-A3B only
- Shared results
- 12
- GLM-5.2 only
- 6
- Qwen3.6-35B-A3B only
- 29
- Like-for-like categories
- 2 / 8
2 categories use different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.
Math
Like-for-like- GLM-5.2
- 95.9
- Qwen3.6-35B-A3B
- 88.2
- Weighted basis
- 2 vs 2 rows
- Reading
- GLM-5.2 leads
Coding
Directional only- GLM-5.2
- 62.1
- Qwen3.6-35B-A3B
- 73.8
- Weighted basis
- 1 vs 3 rows
- Reading
- Directional only
Knowledge
Directional only- GLM-5.2
- 59.6
- Qwen3.6-35B-A3B
- 51.4
- Weighted basis
- 2 vs 4 rows
- Reading
- Directional only
Reasoning
Not comparable- GLM-5.2
- Not measured
- Qwen3.6-35B-A3B
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- GLM-5.2
- Not measured
- Qwen3.6-35B-A3B
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multimodal
Not comparable- GLM-5.2
- Not measured
- Qwen3.6-35B-A3B
- 76.3
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Instruction following
Not comparable- GLM-5.2
- Not measured
- Qwen3.6-35B-A3B
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | GLM-5.2 | Qwen3.6-35B-A3B | Weighted basis | Reading |
|---|---|---|---|---|
| Agentic | 81.0 | 51.5 | Like-for-like1 vs 1 rows | GLM-5.2 leads |
| Math | 95.9 | 88.2 | Like-for-like2 vs 2 rows | GLM-5.2 leads |
| Coding | 62.1 | 73.8 | Directional only1 vs 3 rows | Directional only |
| Knowledge | 59.6 | 51.4 | Directional only2 vs 4 rows | Directional only |
| Reasoning | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multilingual | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multimodal | Not measured | 76.3 | Not comparable0 vs 2 rows | Not comparable |
| Instruction following | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
HLE
Knowledge
Terminal-Bench 2.0
Agentic
SWE-bench Pro
Coding
HMMT Feb 2026
Math
AIME26
Math
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- GLM-5.2
- $0.0036
- Fits in one request
- Qwen3.6-35B-A3B
- API rate not published
- Fits in one request
Qwen3.6-35B-A3B has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- GLM-5.2
- $0.0832
- Fits in one request
- Qwen3.6-35B-A3B
- API rate not published
- Fits in one request
Qwen3.6-35B-A3B has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- GLM-5.2
- $0.352
- Fits in one request
- Cached input priced at the published list-input rate
- Qwen3.6-35B-A3B
- API rate not published
- Fits in one request
- Cached-input rate unavailable
GLM-5.2 has no published cached-input rate, so cached tokens use its listed input rate. Qwen3.6-35B-A3B has no comparable published API token rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationGLM-5.2Qwen3.6-35B-A3B
Context window
Maximum documented context; output-token limits may be lower.
GLM-5.2
1M
Qwen3.6-35B-A3B
262K
API model ID
GLM-5.2
Not sourced
Qwen3.6-35B-A3B
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GLM-5.2
Not published
Qwen3.6-35B-A3B
No comparable hosted API rate
Documented inputs
GLM-5.2
Not sourced
Qwen3.6-35B-A3B
Not sourced
Documented outputs
GLM-5.2
Not sourced
Qwen3.6-35B-A3B
Not sourced
Provider availability
GLM-5.2
Not sourced
Qwen3.6-35B-A3B
Not sourced
Reasoning profile
GLM-5.2
Reasoning
Qwen3.6-35B-A3B
Reasoning
Weight access
GLM-5.2
Open Weight
Qwen3.6-35B-A3B
Open Weight
License
GLM-5.2
Open Weight
Qwen3.6-35B-A3B
Open Weight
Release date
GLM-5.2
2026-06-16
Qwen3.6-35B-A3B
2026-04-15
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- GLM-5.2 has the higher public score estimate, 62.94 versus 50.48, but the 90% score intervals overlap.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- GLM-5.2 has the larger documented window (1M).
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence47 rows
Agentic
Terminal-Bench 2.0
GLM-5.281%SourceQwen3.6-35B-A3B51.5%SourceGLM-5.2 leads this result
MCP Atlas
GLM-5.276.8%SourceQwen3.6-35B-A3B62.8%SourceGLM-5.2 leads this result
Toolathlon
GLM-5.248.2%SourceQwen3.6-35B-A3B26.9%SourceGLM-5.2 leads this result
ResearchClawBench
GLM-5.220.7%SourceQwen3.6-35B-A3B—Not directly comparable
Claw-Eval
GLM-5.2—Qwen3.6-35B-A3B68.7%SourceNot directly comparable
QwenClawBench
GLM-5.2—Qwen3.6-35B-A3B52.6%SourceNot directly comparable
QwenWebBench
GLM-5.2—Qwen3.6-35B-A3B1397SourceNot directly comparable
τ³-bench results
GLM-5.2—Qwen3.6-35B-A3B67.2%SourceNot directly comparable
VITA-Bench
GLM-5.2—Qwen3.6-35B-A3B35.6%SourceNot directly comparable
DeepPlanning
GLM-5.2—Qwen3.6-35B-A3B25.9%SourceNot directly comparable
WideResearch
GLM-5.2—Qwen3.6-35B-A3B60.1%SourceNot directly comparable
Gert Labs
GLM-5.2—Qwen3.6-35B-A3B42.65%SourceNot directly comparable
Coding
SWE-bench Pro
GLM-5.262.1%SourceQwen3.6-35B-A3B49.5%SourceGLM-5.2 leads this result
NL2Repo
GLM-5.248.9%SourceQwen3.6-35B-A3B29.4%SourceGLM-5.2 leads this result
Terminal-Bench 2.0
GLM-5.281.0%SourceQwen3.6-35B-A3B51.5%SourceGLM-5.2 leads this result
ProgramBench
GLM-5.263.7%SourceQwen3.6-35B-A3B—Not directly comparable
cursorBench32
GLM-5.255.0%SourceQwen3.6-35B-A3B—Not directly comparable
SWE-bench Verified
GLM-5.2—Qwen3.6-35B-A3B73.4%SourceNot directly comparable
SWE Multilingual
GLM-5.2—Qwen3.6-35B-A3B67.2%SourceNot directly comparable
LiveCodeBench
GLM-5.2—Qwen3.6-35B-A3B80.4%SourceNot directly comparable
Reasoning
CritPt
GLM-5.220.9%SourceQwen3.6-35B-A3B—Not directly comparable
Knowledge
GPQA
GLM-5.291.2%SourceQwen3.6-35B-A3B86%SourceGLM-5.2 leads this result
GPQA-D
GLM-5.291.2%SourceQwen3.6-35B-A3B—Not directly comparable
HLE
GLM-5.254.7%SourceQwen3.6-35B-A3B21.4%SourceGLM-5.2 leads this result
HLE w/o tools
GLM-5.240.5%SourceQwen3.6-35B-A3B—Not directly comparable
MMLU-Pro
GLM-5.2—Qwen3.6-35B-A3B85.2%SourceNot directly comparable
SuperGPQA
GLM-5.2—Qwen3.6-35B-A3B64.7%SourceNot directly comparable
C-Eval
GLM-5.2—Qwen3.6-35B-A3B90%SourceNot directly comparable
Math
AIME26
GLM-5.299.2%SourceQwen3.6-35B-A3B92.7%SourceGLM-5.2 leads this result
HMMT Nov 2025
GLM-5.294.4%SourceQwen3.6-35B-A3B89.1%SourceGLM-5.2 leads this result
HMMT Feb 2026
GLM-5.292.5%SourceQwen3.6-35B-A3B83.6%SourceGLM-5.2 leads this result
MMAnswerBench
GLM-5.291.0%SourceQwen3.6-35B-A3B78.9%SourceGLM-5.2 leads this result
HMMT Feb 2025
GLM-5.2—Qwen3.6-35B-A3B90.7%SourceNot directly comparable
Multimodal
MMMU
GLM-5.2—Qwen3.6-35B-A3B81.7%SourceNot directly comparable
MMMU-Pro
GLM-5.2—Qwen3.6-35B-A3B75.3%SourceNot directly comparable
RealWorldQA
GLM-5.2—Qwen3.6-35B-A3B85.3%SourceNot directly comparable
OmniDocBench 1.5
GLM-5.2—Qwen3.6-35B-A3B89.9%SourceNot directly comparable
CharXiv
GLM-5.2—Qwen3.6-35B-A3B78%SourceNot directly comparable
SimpleVQA
GLM-5.2—Qwen3.6-35B-A3B58.9%SourceNot directly comparable
CC-OCR
GLM-5.2—Qwen3.6-35B-A3B81.9%SourceNot directly comparable
AI2D_TEST
GLM-5.2—Qwen3.6-35B-A3B92.7%SourceNot directly comparable
RefCOCO (avg)
GLM-5.2—Qwen3.6-35B-A3B92.0%SourceNot directly comparable
ODINW13
GLM-5.2—Qwen3.6-35B-A3B50.8%SourceNot directly comparable
Video-MME (with subtitle)
GLM-5.2—Qwen3.6-35B-A3B86.6%SourceNot directly comparable
Video-MME (w/o subtitle)
GLM-5.2—Qwen3.6-35B-A3B82.5%SourceNot directly comparable
VideoMMMU
GLM-5.2—Qwen3.6-35B-A3B83.7%SourceNot directly comparable
MLVU (M-Avg)
GLM-5.2—Qwen3.6-35B-A3B86.2%SourceNot directly comparable
Frequently asked questions
Which is better, GLM-5.2 or Qwen3.6-35B-A3B?
GLM-5.2 has the higher public score estimate, 62.94 versus 50.48, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, GLM-5.2 or Qwen3.6-35B-A3B?
The current coding averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.
Which is better for agentic tasks, GLM-5.2 or Qwen3.6-35B-A3B?
GLM-5.2 leads the like-for-like agentic tasks comparison across 1 shared weighted benchmark row.
Which costs less, GLM-5.2 or Qwen3.6-35B-A3B?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, GLM-5.2 or Qwen3.6-35B-A3B?
GLM-5.2 has the larger documented context window: 1M, compared with 262K.
Related comparisons
Last updated July 31, 2026
Watch GLM-5.2 vs Qwen3.6-35B-A3B
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.