Agentic
Directional only- Claude Sonnet 4.5
- 55.4
- GLM-5
- 56.2
- Weighted basis
- 2 vs 1 rows
- Reading
- Directional only
Claude Sonnet 4.5 vs GLM-5
Updated September 3, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
GLM-5 has the higher public score estimate, 65.68 versus 54.88, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
6 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.
Chat turn cost
1K fresh input + 500 output tokens
GLM-5
GLM-5 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Show secondary and unsupported calls
Repository review cost
50K fresh input + 3K output tokens
GLM-5
GLM-5 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
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
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
Long documents
Prompts that approach the documented context limit
No clear pick
The documented context windows are equal.
Confidence: documented
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. Claude Sonnet 4.5 does not fit this workload in one request. GLM-5 does not fit this workload in one request. Claude Sonnet 4.5 has no published cached-input rate, so cached tokens use its listed input rate. GLM-5 has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallback
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
5 Claude Sonnet 4.5 only
6 Shared
30 GLM-5 only
- Shared results
- 6
- Claude Sonnet 4.5 only
- 5
- GLM-5 only
- 30
- Like-for-like categories
- 0 / 8
4 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.
Coding
Directional only- Claude Sonnet 4.5
- 77.2
- GLM-5
- 66.3
- Weighted basis
- 1 vs 3 rows
- Reading
- Directional only
Knowledge
Directional only- Claude Sonnet 4.5
- 83.4
- GLM-5
- 66.4
- Weighted basis
- 1 vs 4 rows
- Reading
- Directional only
Math
Directional only- Claude Sonnet 4.5
- 11.2
- GLM-5
- 56.3
- Weighted basis
- 2 vs 4 rows
- Reading
- Directional only
Reasoning
Not comparable- Claude Sonnet 4.5
- 13.6
- GLM-5
- 60.8
- Weighted basis
- 1 vs 1 rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Sonnet 4.5
- Not measured
- GLM-5
- 83.1
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Multimodal
Not comparable- Claude Sonnet 4.5
- Not measured
- GLM-5
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Sonnet 4.5
- Not measured
- GLM-5
- 92.6
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
| Category | Claude Sonnet 4.5 | GLM-5 | Weighted basis | Reading |
|---|---|---|---|---|
| Agentic | 55.4 | 56.2 | Directional only2 vs 1 rows | Directional only |
| Coding | 77.2 | 66.3 | Directional only1 vs 3 rows | Directional only |
| Knowledge | 83.4 | 66.4 | Directional only1 vs 4 rows | Directional only |
| Math | 11.2 | 56.3 | Directional only2 vs 4 rows | Directional only |
| Reasoning | 13.6 | 60.8 | Not comparable1 vs 1 rows | Not comparable |
| Multilingual | Not measured | 83.1 | Not comparable0 vs 1 rows | Not comparable |
| Multimodal | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Instruction following | Not measured | 92.6 | Not comparable0 vs 1 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.
Terminal-Bench 2.0
Agentic
FrontierMath v2 (Tiers 1-3)
Math
GPQA
Knowledge
FrontierMath v2 (Tier 4)
Math
SWE-bench Verified
Coding
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
- Claude Sonnet 4.5
- $0.0105
- Fits in one request
- GLM-5
- $0.0026
- Fits in one request
GLM-5 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Sonnet 4.5
- $0.195
- Fits in one request
- GLM-5
- $0.0596
- Fits in one request
GLM-5 has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Claude Sonnet 4.5
- $0.81
- Does not fit in one request
- Cached input priced at the published list-input rate
- GLM-5
- $0.252
- Does not fit in one request
- Cached input priced at the published list-input rate
Claude Sonnet 4.5 does not fit this workload in one request. GLM-5 does not fit this workload in one request. Claude Sonnet 4.5 has no published cached-input rate, so cached tokens use its listed input rate. GLM-5 has no published cached-input rate, so cached tokens use its listed input rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationClaude Sonnet 4.5GLM-5
Context window
Maximum documented context; output-token limits may be lower.
Claude Sonnet 4.5
200K
GLM-5
200K
API model ID
Claude Sonnet 4.5
Not sourced
GLM-5
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Sonnet 4.5
Not published
GLM-5
Not published
Documented inputs
Claude Sonnet 4.5
Not sourced
GLM-5
Not sourced
Documented outputs
Claude Sonnet 4.5
Not sourced
GLM-5
Not sourced
Provider availability
Claude Sonnet 4.5
Not sourced
GLM-5
Not sourced
Reasoning profile
Claude Sonnet 4.5
Non-Reasoning
GLM-5
Non-Reasoning
Weight access
Claude Sonnet 4.5
Proprietary
GLM-5
Open Weight
License
Claude Sonnet 4.5
Proprietary
GLM-5
Open Weight
Release date
Claude Sonnet 4.5
2025-09-01
GLM-5
2026-03-01
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 has the higher public score estimate, 65.68 versus 54.88, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.195 vs $0.0596. Cache-heavy agent loop: $0.81 vs $0.252.
- Context tradeoff
- Both models list 200K.
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 evidence41 rows
Agentic
Terminal-Bench 2.0
Claude Sonnet 4.550%SourceGLM-556.2%SourceGLM-5 leads this result
OSWorld-Verified
Claude Sonnet 4.561.4%SourceGLM-5—Not directly comparable
VITA-Bench
Claude Sonnet 4.517.0%SourceGLM-5—Not directly comparable
Gert Labs
Shared sourceClaude Sonnet 4.548.51%GLM-550.99%GLM-5 leads this result
JobBench
Claude Sonnet 4.527.7%SourceGLM-5—Not directly comparable
Claw-Eval
Claude Sonnet 4.5—GLM-557.7%SourceNot directly comparable
QwenClawBench
Claude Sonnet 4.5—GLM-554.1%SourceNot directly comparable
τ³-bench results
Claude Sonnet 4.5—GLM-565.6%SourceNot directly comparable
DeepPlanning
Claude Sonnet 4.5—GLM-514.6%SourceNot directly comparable
Toolathlon
Claude Sonnet 4.5—GLM-538%SourceNot directly comparable
MCP Atlas
Claude Sonnet 4.5—GLM-531.1%SourceNot directly comparable
MCP-Tasks
Claude Sonnet 4.5—GLM-560.8%SourceNot directly comparable
WideResearch
Claude Sonnet 4.5—GLM-569.8%SourceNot directly comparable
CyberGym
Claude Sonnet 4.5—GLM-543.2%SourceNot directly comparable
Coding
SWE-bench Verified
Claude Sonnet 4.577.2%SourceGLM-577.8%SourceGLM-5 leads this result
SWE-bench Verified*
Claude Sonnet 4.5—GLM-572.8%SourceNot directly comparable
SWE-bench Pro
Claude Sonnet 4.5—GLM-555.1%SourceNot directly comparable
SWE Multilingual
Claude Sonnet 4.5—GLM-573.3%SourceNot directly comparable
SWE-Rebench
Claude Sonnet 4.5—GLM-562.8%SourceNot directly comparable
React Native Evals
Claude Sonnet 4.5—GLM-574.8%SourceNot directly comparable
Reasoning
Knowledge
GPQA
Claude Sonnet 4.583.4%SourceGLM-586%SourceGLM-5 leads this result
GPQA-D
Claude Sonnet 4.5—GLM-586.0%SourceNot directly comparable
SuperGPQA
Claude Sonnet 4.5—GLM-566.8%SourceNot directly comparable
MMLU-Pro
Claude Sonnet 4.5—GLM-585.7%SourceNot directly comparable
MMLU-Pro (Arcee)
Claude Sonnet 4.5—GLM-585.8%SourceNot directly comparable
HLE
Claude Sonnet 4.5—GLM-550.4%SourceNot directly comparable
Math
AIME 2025
Claude Sonnet 4.587%SourceGLM-5—Not directly comparable
FrontierMath v2 (Tiers 1-3)
Shared sourceClaude Sonnet 4.513.495%GLM-516.434%GLM-5 leads this result
FrontierMath v2 (Tier 4)
Shared sourceClaude Sonnet 4.54.167%GLM-52.100%Claude Sonnet 4.5 leads this result
AIME26
Claude Sonnet 4.5—GLM-595.8%SourceNot directly comparable
AIME25 (Arcee)
Claude Sonnet 4.5—GLM-593.3%SourceNot directly comparable
HMMT Feb 2025
Claude Sonnet 4.5—GLM-597.5%SourceNot directly comparable
HMMT Nov 2025
Claude Sonnet 4.5—GLM-596.9%SourceNot directly comparable
HMMT Feb 2026
Claude Sonnet 4.5—GLM-586.4%SourceNot directly comparable
MMAnswerBench
Claude Sonnet 4.5—GLM-582.5%SourceNot directly comparable
Multilingual
Instruction following
IFEval
Claude Sonnet 4.5—GLM-592.6%SourceNot directly comparable
Frequently asked questions
Which is better, Claude Sonnet 4.5 or GLM-5?
GLM-5 has the higher public score estimate, 65.68 versus 54.88, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, Claude Sonnet 4.5 or GLM-5?
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, Claude Sonnet 4.5 or GLM-5?
The current agentic tasks 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 costs less, Claude Sonnet 4.5 or GLM-5?
For the stated presets, chat costs $0.0105 on Claude Sonnet 4.5 and $0.0026 on GLM-5; repository review costs $0.195 and $0.0596; the cache-heavy agent loop costs $0.81 and $0.252. Claude Sonnet 4.5 does not fit this workload in one request. GLM-5 does not fit this workload in one request. Claude Sonnet 4.5 has no published cached-input rate, so cached tokens use its listed input rate. GLM-5 has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, Claude Sonnet 4.5 or GLM-5?
Both models list the same context window, 200K.
Related comparisons
Last updated September 3, 2026
Watch Claude Sonnet 4.5 vs GLM-5
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.