Coding
Like-for-like- Claude Opus 4.8
- 81.1
- Claude Opus 5
- 89.5
- Weighted basis
- 2 vs 2 rows
- Reading
- Claude Opus 5 leads
Claude Opus 4.8 vs Claude Opus 5
Updated August 22, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Claude Opus 5 has the higher public score, 82.72 versus 76.2, and the 90% score intervals do not overlap.
17 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.
Coding work
Code generation, repair, and software-engineering tasks
Claude Opus 5
Claude Opus 5 leads on the same 2 weighted benchmark rows.
Confidence: limited
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Claude Opus 5
Claude Opus 5 has the lower estimated token cost for this stated workload. Claude Opus 4.8 has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallback
Show secondary and unsupported calls
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
Chat turn cost
1K fresh input + 500 output tokens
No clear pick
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
No clear pick
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.
16 Claude Opus 4.8 only
17 Shared
46 Claude Opus 5 only
- Shared results
- 17
- Claude Opus 4.8 only
- 16
- Claude Opus 5 only
- 46
- Like-for-like categories
- 2 / 8
3 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.
Reasoning
Like-for-like- Claude Opus 4.8
- 72.1
- Claude Opus 5
- 90.4
- Weighted basis
- 1 vs 1 rows
- Reading
- Claude Opus 5 leads
Agentic
Directional only- Claude Opus 4.8
- 80.3
- Claude Opus 5
- 90.8
- Weighted basis
- 3 vs 1 rows
- Reading
- Directional only
Knowledge
Directional only- Claude Opus 4.8
- 62.7
- Claude Opus 5
- 64.7
- Weighted basis
- 2 vs 1 rows
- Reading
- Directional only
Multimodal
Directional only- Claude Opus 4.8
- 77.0
- Claude Opus 5
- 66.9
- Weighted basis
- 2 vs 1 rows
- Reading
- Directional only
Math
Not comparable- Claude Opus 4.8
- 53.9
- Claude Opus 5
- Not measured
- Weighted basis
- 3 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Opus 4.8
- Not measured
- Claude Opus 5
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Opus 4.8
- Not measured
- Claude Opus 5
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | Claude Opus 4.8 | Claude Opus 5 | Weighted basis | Reading |
|---|---|---|---|---|
| Coding | 81.1 | 89.5 | Like-for-like2 vs 2 rows | Claude Opus 5 leads |
| Reasoning | 72.1 | 90.4 | Like-for-like1 vs 1 rows | Claude Opus 5 leads |
| Agentic | 80.3 | 90.8 | Directional only3 vs 1 rows | Directional only |
| Knowledge | 62.7 | 64.7 | Directional only2 vs 1 rows | Directional only |
| Multimodal | 77.0 | 66.9 | Directional only2 vs 1 rows | Directional only |
| Math | 53.9 | Not measured | Not comparable3 vs 0 rows | Not comparable |
| Multilingual | Not measured | Not measured | Not comparable0 vs 0 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.
ARC-AGI-2
Reasoning
Claude Opus 4.8: 72.1%Claude Opus 5: 90.4%Normalized gap 18.3Claude Opus 4.8 source Claude Opus 5 sourceSWE-bench Pro
Coding
Claude Opus 4.8: 69.2%Claude Opus 5: 79.2%Normalized gap 10.0Claude Opus 4.8 source Claude Opus 5 sourceSWE-bench Verified
Coding
Claude Opus 4.8: 88.6%Claude Opus 5: 96%Normalized gap 7.4Claude Opus 4.8 source Claude Opus 5 sourceHLE
Knowledge
Claude Opus 4.8: 57.9%Claude Opus 5: 64.7%Normalized gap 6.8Claude Opus 4.8 source Claude Opus 5 sourceBrowseComp
Agentic
Claude Opus 4.8: 84.3%Claude Opus 5: 90.8%Normalized gap 6.5Claude Opus 4.8 source Claude Opus 5 source
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 Opus 4.8
- $0.0175
- Fits in one request
- Claude Opus 5
- $0.0175
- Fits in one request
Modeled costs are equal
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Opus 4.8
- $0.325
- Fits in one request
- Claude Opus 5
- $0.325
- Fits in one request
Modeled costs are equal
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Claude Opus 4.8
- $1.35
- Fits in one request
- Cached input priced at the published list-input rate
- Claude Opus 5
- $0.45
- Fits in one request
Claude Opus 5 has the lower modeled cost
Claude Opus 4.8 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 Opus 4.8Claude Opus 5
Context window
Maximum documented context; output-token limits may be lower.
Claude Opus 4.8
Claude Opus 5
API model ID
Claude Opus 4.8
claude-opus-4-8
Anthropic model overviewClaude Opus 5
claude-opus-5
Anthropic model overviewCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Opus 4.8
Not published
Claude Opus 5
$0.5 per 1M cached input tokens
Claude API pricingDocumented inputs
Claude Opus 4.8
text, image
Anthropic model overviewClaude Opus 5
text, image
Anthropic model overviewDocumented outputs
Claude Opus 4.8
Claude Opus 5
Provider availability
Claude Opus 4.8
Generally Available · Claude API
Anthropic model overviewClaude Opus 5
Generally Available · Claude API
Anthropic model overviewReasoning profile
Claude Opus 4.8
Reasoning
Claude Opus 5
Reasoning
Weight access
Claude Opus 4.8
Proprietary
Claude Opus 5
Proprietary
License
Claude Opus 4.8
Proprietary
Claude Opus 5
Proprietary
Release date
Claude Opus 4.8
2026-05-28
Claude Opus 5
2026-07-24
If you are considering the documented upgrade path
- Deployment change
- Both entries list Anthropic as the provider. Confirm endpoint, model ID, limits, and feature support before switching.
- Quality signal
- Claude Opus 5 has the higher public score, 82.72 versus 76.2, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.325 vs $0.325. Cache-heavy agent loop: $1.35 vs $0.45.
- Context tradeoff
- Both models list 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 evidence79 rows
Agentic
Terminal-Bench 3.0
Shared sourceClaude Opus 4.821.1%Claude Opus 542.7%Claude Opus 5 leads this result
Terminal-Bench 2.0
Claude Opus 4.874.6%SourceClaude Opus 5—Not directly comparable
BrowseComp
Claude Opus 4.884.3%SourceClaude Opus 590.8%SourceClaude Opus 5 leads this result
DeepSearchQA
Claude Opus 4.893.1%SourceClaude Opus 595.0%SourceClaude Opus 5 leads this result
OSWorld-Verified
Claude Opus 4.883.4%SourceClaude Opus 5—Not directly comparable
Finance Agent v2
Claude Opus 4.853.9%SourceClaude Opus 5—Not directly comparable
MCP Atlas
Claude Opus 4.882.2%SourceClaude Opus 585.8%SourceClaude Opus 5 leads this result
Toolathlon
Claude Opus 4.859.9%SourceClaude Opus 5—Not directly comparable
Gert Labs
Claude Opus 4.872.97%SourceClaude Opus 5—Not directly comparable
ResearchClawBench
Claude Opus 4.821.1%SourceClaude Opus 5—Not directly comparable
OSWorld 2.0
Claude Opus 4.820.6%SourceClaude Opus 570.6%SourceClaude Opus 5 leads this result
HLE w/ tools
Claude Opus 4.8—Claude Opus 564.7%SourceNot directly comparable
DRACO
Claude Opus 4.8—Claude Opus 588.6%SourceNot directly comparable
BrowseComp (10-agent, prerelease)
Claude Opus 4.8—Claude Opus 593.6%SourceNot directly comparable
MCP-Atlas claim coverage
Claude Opus 4.8—Claude Opus 589.1%SourceNot directly comparable
LAB all-pass (Anthropic harness)
Claude Opus 4.8—Claude Opus 523.58%SourceNot directly comparable
LAB criterion-pass (Anthropic harness)
Claude Opus 4.8—Claude Opus 593.74%SourceNot directly comparable
LAB all-pass (Harvey held-out)
Claude Opus 4.8—Claude Opus 511.7%SourceNot directly comparable
LAB criterion-pass (Harvey held-out)
Claude Opus 4.8—Claude Opus 594.1%SourceNot directly comparable
Toolathlon-Verified
Claude Opus 4.8—Claude Opus 580.6%SourceNot directly comparable
Toolathlon Verified Pass@3
Claude Opus 4.8—Claude Opus 587.0%SourceNot directly comparable
Toolathlon Verified Pass³
Claude Opus 4.8—Claude Opus 573.1%SourceNot directly comparable
Toolathlon Verified avg. turns
Claude Opus 4.8—Claude Opus 523.5 turnsSourceNot directly comparable
AutomationBench
Claude Opus 4.8—Claude Opus 526.0%SourceNot directly comparable
Coding
SWE-bench Verified
Claude Opus 4.888.6%SourceClaude Opus 596%SourceClaude Opus 5 leads this result
SWE-bench Pro
Claude Opus 4.869.2%SourceClaude Opus 579.2%SourceClaude Opus 5 leads this result
SWE Multilingual
Claude Opus 4.884.4%SourceClaude Opus 589.5%SourceClaude Opus 5 leads this result
SWE Multimodal
Claude Opus 4.838.4%SourceClaude Opus 559.4%SourceClaude Opus 5 leads this result
Terminal-Bench 2.0
Claude Opus 4.874.6%SourceClaude Opus 5—Not directly comparable
cursorBench31
Claude Opus 4.858.4%SourceClaude Opus 5—Not directly comparable
cursorBench32
Shared sourceClaude Opus 4.862.3%Claude Opus 570.0%Claude Opus 5 leads this result
FrontierCode 1.1 Main
Claude Opus 4.846.5%SourceClaude Opus 553.4%SourceClaude Opus 5 leads this result
deepSwe
Claude Opus 4.8—Claude Opus 568.8%SourceNot directly comparable
FrontierCode 1.1 Extended
Claude Opus 4.8—Claude Opus 563.6%SourceNot directly comparable
ProgramBench (episode 1)
Claude Opus 4.8—Claude Opus 583.0%SourceNot directly comparable
ProgramBench
Claude Opus 4.8—Claude Opus 593.0%SourceNot directly comparable
VulcanBench v3
Claude Opus 4.8—Claude Opus 587.0%SourceNot directly comparable
Reasoning
Knowledge
GPQA
Claude Opus 4.893.6%SourceClaude Opus 5—Not directly comparable
GPQA-D
Claude Opus 4.893.6%SourceClaude Opus 5—Not directly comparable
HLE
Claude Opus 4.857.9%SourceClaude Opus 564.7%SourceClaude Opus 5 leads this result
HLE w/o tools
Claude Opus 4.849.8%SourceClaude Opus 556.3%SourceClaude Opus 5 leads this result
HealthBench (raw)
Claude Opus 4.8—Claude Opus 567.1%SourceNot directly comparable
HealthBench (length-adjusted)
Claude Opus 4.8—Claude Opus 557.8%SourceNot directly comparable
HealthBench Professional
Claude Opus 4.8—Claude Opus 559.8%SourceNot directly comparable
HealthBench Professional (raw)
Claude Opus 4.8—Claude Opus 573.4%SourceNot directly comparable
BioMysteryBench (human-solvable)
Claude Opus 4.8—Claude Opus 590.1%SourceNot directly comparable
BioMysteryBench (human-difficult)
Claude Opus 4.8—Claude Opus 549.4%SourceNot directly comparable
SpatialBench Verified
Claude Opus 4.8—Claude Opus 572.5%SourceNot directly comparable
SingleCellBench
Claude Opus 4.8—Claude Opus 560.6%SourceNot directly comparable
ProteinGym Hard
Claude Opus 4.8—Claude Opus 547.7%SourceNot directly comparable
Protein Design
Claude Opus 4.8—Claude Opus 542.5%SourceNot directly comparable
Organic chemistry V2
Claude Opus 4.8—Claude Opus 561.6%SourceNot directly comparable
Protocols (troubleshooting)
Claude Opus 4.8—Claude Opus 561.1%SourceNot directly comparable
Protocols (understanding)
Claude Opus 4.8—Claude Opus 578.4%SourceNot directly comparable
Math
USAMO 2026
Claude Opus 4.896.7%SourceClaude Opus 5—Not directly comparable
FrontierMath v2 (Tiers 1-3)
Claude Opus 4.847.241%SourceClaude Opus 5—Not directly comparable
FrontierMath v2 (Tier 4)
Claude Opus 4.831.250%SourceClaude Opus 5—Not directly comparable
IMO 2026
Claude Opus 4.8—Claude Opus 542/42SourceNot directly comparable
RiemannBench (no tools)
Claude Opus 4.8—Claude Opus 560.0%SourceNot directly comparable
RiemannBench (tools)
Claude Opus 4.8—Claude Opus 579.0%SourceNot directly comparable
ArXivMath Jun. 2026 (no tools)
Claude Opus 4.8—Claude Opus 590.8%SourceNot directly comparable
ArXivMath Jun. 2026 (tools)
Claude Opus 4.8—Claude Opus 591.3%SourceNot directly comparable
Multilingual
Multimodal
OfficeQA Pro
Claude Opus 4.866.2%SourceClaude Opus 566.9%SourceClaude Opus 5 leads this result
ScreenSpot Pro
Claude Opus 4.887.9%SourceClaude Opus 5—Not directly comparable
CharXiv
Claude Opus 4.889.9%SourceClaude Opus 5—Not directly comparable
CharXiv w/o tools
Claude Opus 4.880.5%SourceClaude Opus 5—Not directly comparable
Chartography (no tools)
Claude Opus 4.8—Claude Opus 529.6%SourceNot directly comparable
Chartography (tools)
Claude Opus 4.8—Claude Opus 583.0%SourceNot directly comparable
BenchCAD Vision2Code (no tools)
Claude Opus 4.8—Claude Opus 50.366SourceNot directly comparable
BenchCAD Vision2Code (tools)
Claude Opus 4.8—Claude Opus 50.821SourceNot directly comparable
GDP.pdf (no tools)
Claude Opus 4.8—Claude Opus 583.4%SourceNot directly comparable
GDP.pdf (tools)
Claude Opus 4.8—Claude Opus 585.5%SourceNot directly comparable
OfficeQA
Claude Opus 4.8—Claude Opus 578.1%SourceNot directly comparable
Frequently asked questions
Which is better, Claude Opus 4.8 or Claude Opus 5?
Claude Opus 5 has the higher public score, 82.72 versus 76.2, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.
Which is better for coding, Claude Opus 4.8 or Claude Opus 5?
Claude Opus 5 leads the like-for-like coding comparison across 2 shared weighted benchmark rows.
Which is better for agentic tasks, Claude Opus 4.8 or Claude Opus 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 Opus 4.8 or Claude Opus 5?
For the stated presets, chat costs $0.0175 on Claude Opus 4.8 and $0.0175 on Claude Opus 5; repository review costs $0.325 and $0.325; the cache-heavy agent loop costs $1.35 and $0.45. Claude Opus 4.8 has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, Claude Opus 4.8 or Claude Opus 5?
Both models list the same context window, 1M.
Related comparisons
Last updated August 22, 2026
Watch Claude Opus 4.8 vs Claude Opus 5
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.