Coding work
Code generation, repair, and software-engineering tasks
Claude Opus 5
Claude Opus 5 leads on the public coding lane, 72.8 to 62, with Supported evidence for both models, although the 90% intervals overlap.
Confidence: limited
Claude Opus 5 vs Grok 4.6
Updated September 25, 2026. Rank says Claude Opus 5 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Both of these models will change. Get the price, version and retirement notices for the pair, each with its source. Follow model changes
Share or export
Decision reading
Claude Opus 5 has the higher public score, 79.81 versus 69.19, and the 90% score intervals do not overlap. 16 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 16
- Claude Opus 5 only
- 58
- Grok 4.6 only
- 1
- Like-for-like categories
- 4 / 8
Supported: Claude Opus 5 and Grok 4.6How the comparison works
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
Claude Opus 5
Claude Opus 5 leads on the public agentic lane, 77.4 to 67.9, with Supported evidence for both models, although the 90% intervals overlap.
Confidence: limitedLong documents
Prompts that approach the documented context limit
Claude Opus 5
Claude Opus 5 has the larger documented context window.
Confidence: documented
Show secondary and unsupported calls
Chat turn cost
1K fresh input + 500 output tokens
Grok 4.6
Grok 4.6 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Grok 4.6
Grok 4.6 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesRepository review cost
50K fresh input + 3K output tokens
Grok 4.6
Grok 4.6 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Which one for a specific job
Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
72.8Claude Opus 562.0Grok 4.6
Like-for-like · BenchAlign v5.7
Claude Opus 5 leads the like-for-like coding row, although the 90% intervals overlap.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
Same basis rules as the category table below
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
58 Claude Opus 5 only
16 Shared
1 Grok 4.6 only
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-3Reasoning
Normalized gap 28.0Claude Opus 5:30.2%Grok 4.6:2.1%ARC-AGI-2Reasoning
Normalized gap 23.3Claude Opus 5:90.4%Grok 4.6:67.1%MMLU-Pro (Vals)Knowledge
Normalized gap 2.2Claude Opus 5:91.6%Grok 4.6:89.4%cursorBench32Coding
Normalized gap 0.8Claude Opus 5:70.0%Grok 4.6:70.8%LiveCodeBench (Vals)Coding
Normalized gap 0.8Claude Opus 5:89.0%Grok 4.6:88.2%
Bars run 0–100 on each benchmark’s normalized display scale
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign v5.7 lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
Agentic
Like-for-like- Claude Opus 5
- 77.4
- Supported · #3/105
- Grok 4.6
- 67.9
- Supported · #8/105
- Basis
- BenchAlign v5.7 lane · 20 vs 4 public rows
- Reading
- Claude Opus 5 leads · intervals overlap
Coding
Like-for-like- Claude Opus 5
- 72.8
- Supported · #5/135
- Grok 4.6
- 62.0
- Supported · #14/135
- Basis
- BenchAlign v5.7 lane · 16 vs 8 public rows
- Reading
- Claude Opus 5 leads · intervals overlap
Reasoning
Like-for-like- Claude Opus 5
- 77.3
- #6/19
- Grok 4.6
- 57.6
- #16/19
- Basis
- Provisional lane · 2 vs 2 weighted rows
- Reading
- Claude Opus 5 leads
Knowledge
Like-for-like- Claude Opus 5
- 81.4
- Supported · #5/158
- Grok 4.6
- 69.0
- Supported · #13/158
- Basis
- BenchAlign v5.7 lane · 19 vs 2 public rows
- Reading
- Claude Opus 5 leads · intervals overlap
Multimodal
Not comparable- Claude Opus 5
- 88.8
- #2/50
- Grok 4.6
- Not ranked
- Basis
- Provisional lane · 1 vs 0 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Opus 5
- Not ranked
- Grok 4.6
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Opus 5
- Not ranked
- Grok 4.6
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- Claude Opus 5
- Not ranked
- Grok 4.6
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | Claude Opus 5 | Grok 4.6 | Basis | Reading |
|---|---|---|---|---|
| Agentic | 77.4Supported · #3/105 | 67.9Supported · #8/105 | Like-for-likeBenchAlign v5.7 lane · 20 vs 4 public rows | Claude Opus 5 leads · intervals overlap |
| Coding | 72.8Supported · #5/135 | 62.0Supported · #14/135 | Like-for-likeBenchAlign v5.7 lane · 16 vs 8 public rows | Claude Opus 5 leads · intervals overlap |
| Reasoning | 77.3#6/19 | 57.6#16/19 | Like-for-likeProvisional lane · 2 vs 2 weighted rows | Claude Opus 5 leads |
| Knowledge | 81.4Supported · #5/158 | 69.0Supported · #13/158 | Like-for-likeBenchAlign v5.7 lane · 19 vs 2 public rows | Claude Opus 5 leads · intervals overlap |
| Multimodal | 88.8#2/50 | Not ranked | Not comparableProvisional lane · 1 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.7) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Supported evidence per lane · bars run 0–100Methodology
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 5
- $0.0175
- Fits in one request
- Grok 4.6
- $0.005
- Fits in one request
Grok 4.6 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Opus 5
- $0.325
- Fits in one request
- Grok 4.6
- $0.118
- Fits in one request
Grok 4.6 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 Opus 5
- $0.45
- Fits in one request
- Grok 4.6
- $0.2
- Fits in one request
Grok 4.6 has the lower modeled cost
Costs use the listed standard API rates.
Cached input falls back to the list input rate only where a cached rate is unpublished
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationClaude Opus 5Grok 4.6
Context window
Maximum documented context; output-token limits may be lower.
Claude Opus 5
Grok 4.6
API model ID
Claude Opus 5
claude-opus-5
Anthropic model overviewGrok 4.6
grok-4.6
xAI Grok 4.6 release notesCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Opus 5
$0.5 per 1M cached input tokens
Claude API pricingGrok 4.6
$0.5 per 1M cached input tokens
xAI Grok 4.6 release notesDocumented inputs
Claude Opus 5
text, image
Anthropic model overviewGrok 4.6
Not sourced
Documented outputs
Claude Opus 5
Grok 4.6
Not sourced
Provider availability
Claude Opus 5
Generally Available · Claude API
Anthropic model overviewGrok 4.6
Not sourced
Reasoning profile
Claude Opus 5
Reasoning
Grok 4.6
Reasoning
Weight access
Claude Opus 5
Proprietary
Grok 4.6
Proprietary
License
Claude Opus 5
Proprietary
Grok 4.6
Proprietary
Release date
Claude Opus 5
2026-07-24
Grok 4.6
2026-08-12
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
- Claude Opus 5 has the higher public score, 79.81 versus 69.19, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.325 vs $0.118. Cache-heavy agent loop: $0.45 vs $0.2.
- Context tradeoff
- Claude Opus 5 has the larger documented window (1M).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, Claude Opus 5 or Grok 4.6?
Claude Opus 5 has the higher public score, 79.81 versus 69.19, 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 5 or Grok 4.6?
Claude Opus 5 leads the public coding lane, 72.8 to 62, with Supported evidence for both models, although the 90% intervals overlap.
Which is better for agentic tasks, Claude Opus 5 or Grok 4.6?
Claude Opus 5 leads the public agentic tasks lane, 77.4 to 67.9, with Supported evidence for both models, although the 90% intervals overlap.
Which costs less, Claude Opus 5 or Grok 4.6?
For the stated presets, chat costs $0.0175 on Claude Opus 5 and $0.005 on Grok 4.6; repository review costs $0.325 and $0.118; the cache-heavy agent loop costs $0.45 and $0.2. Costs use the listed standard API rates.
Which has the larger context window, Claude Opus 5 or Grok 4.6?
Claude Opus 5 has the larger documented context window: 1M, compared with 500K.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence75 rows
Agentic
Terminal-Bench 3.0
Shared sourceClaude Opus 542.7%Grok 4.626.5%Claude Opus 5 leads this result
BrowseComp
Claude Opus 590.8%SourceGrok 4.6—Not directly comparable
HLE w/ tools
Claude Opus 564.7%SourceGrok 4.6—Not directly comparable
DeepSearchQA
Claude Opus 595.0%SourceGrok 4.6—Not directly comparable
DRACO
Claude Opus 588.6%SourceGrok 4.6—Not directly comparable
BrowseComp (10-agent, prerelease)
Claude Opus 593.6%SourceGrok 4.6—Not directly comparable
OSWorld 2.0
Claude Opus 570.6%SourceGrok 4.6—Not directly comparable
MCP Atlas
Claude Opus 585.8%SourceGrok 4.6—Not directly comparable
MCP-Atlas claim coverage
Claude Opus 589.1%SourceGrok 4.6—Not directly comparable
LAB all-pass (Anthropic harness)
Claude Opus 523.58%SourceGrok 4.6—Not directly comparable
LAB criterion-pass (Anthropic harness)
Claude Opus 593.74%SourceGrok 4.6—Not directly comparable
LAB all-pass (Harvey held-out)
Claude Opus 511.7%SourceGrok 4.6—Not directly comparable
LAB criterion-pass (Harvey held-out)
Claude Opus 594.1%SourceGrok 4.6—Not directly comparable
Toolathlon-Verified
Claude Opus 580.6%SourceGrok 4.6—Not directly comparable
Toolathlon Verified Pass@3
Claude Opus 587.0%SourceGrok 4.6—Not directly comparable
Toolathlon Verified Pass³
Claude Opus 573.1%SourceGrok 4.6—Not directly comparable
Toolathlon Verified avg. turns
Claude Opus 523.5 turnsSourceGrok 4.6—Not directly comparable
AutomationBench
Claude Opus 526.0%SourceGrok 4.6—Not directly comparable
Terminal-Bench 2.1 (Vals)
Claude Opus 584.6%SourceGrok 4.678.3%SourceClaude Opus 5 leads this result
ApprenticeBench
Shared sourceClaude Opus 536%Grok 4.613%Claude Opus 5 leads this result
APEX-Agents
Claude Opus 5—Grok 4.657.5%SourceNot directly comparable
Coding
Bug Hunt Bench
Shared sourceClaude Opus 527 fixesGrok 4.627 fixesTie
SWE-bench Verified
Claude Opus 596%SourceGrok 4.6—Not directly comparable
SWE-bench Pro
Claude Opus 579.2%SourceGrok 4.6—Not directly comparable
SWE Multilingual
Claude Opus 589.5%SourceGrok 4.6—Not directly comparable
SWE Multimodal
Claude Opus 559.4%SourceGrok 4.6—Not directly comparable
DeepSWE
Claude Opus 568.8%SourceGrok 4.665.9%SourceClaude Opus 5 leads this result
FrontierCode 1.1 Main
Claude Opus 553.4%SourceGrok 4.6—Not directly comparable
FrontierCode 1.1 Extended
Claude Opus 563.6%SourceGrok 4.661.3%SourceClaude Opus 5 leads this result
FrontierSWE v2
Shared sourceClaude Opus 552.0%Grok 4.625.3%Claude Opus 5 leads this result
ProgramBench (episode 1)
Claude Opus 583.0%SourceGrok 4.6—Not directly comparable
ProgramBench
Claude Opus 593.0%SourceGrok 4.6—Not directly comparable
cursorBench32
Shared sourceClaude Opus 570.0%Grok 4.670.8%Grok 4.6 leads this result
VulcanBench v3
Claude Opus 587.0%SourceGrok 4.687.0%SourceTie
VulcanBench CII v1
Claude Opus 596.4%SourceGrok 4.6—Not directly comparable
LiveCodeBench (Vals)
Claude Opus 589.0%SourceGrok 4.688.2%SourceClaude Opus 5 leads this result
SWE-bench (Vals)
Claude Opus 597.0%SourceGrok 4.695.6%SourceClaude Opus 5 leads this result
Reasoning
Multimodal
Chartography (no tools)
Claude Opus 529.6%SourceGrok 4.6—Not directly comparable
Chartography (tools)
Claude Opus 583.0%SourceGrok 4.6—Not directly comparable
BenchCAD Vision2Code (no tools)
Claude Opus 50.366SourceGrok 4.6—Not directly comparable
BenchCAD Vision2Code (tools)
Claude Opus 50.821SourceGrok 4.6—Not directly comparable
GDP.pdf (no tools)
Claude Opus 583.4%SourceGrok 4.6—Not directly comparable
GDP.pdf (tools)
Claude Opus 585.5%SourceGrok 4.6—Not directly comparable
OfficeQA
Claude Opus 578.1%SourceGrok 4.6—Not directly comparable
OfficeQA Pro
Claude Opus 566.9%SourceGrok 4.6—Not directly comparable
Knowledge
HLE
Claude Opus 564.7%SourceGrok 4.6—Not directly comparable
HLE w/o tools
Claude Opus 556.3%SourceGrok 4.6—Not directly comparable
HLE-Verified
Claude Opus 554.4%SourceGrok 4.6—Not directly comparable
LABBench2
Claude Opus 584.2%SourceGrok 4.6—Not directly comparable
HealthBench (raw)
Claude Opus 567.1%SourceGrok 4.6—Not directly comparable
HealthBench (length-adjusted)
Claude Opus 557.8%SourceGrok 4.6—Not directly comparable
HealthBench Professional
Claude Opus 559.8%SourceGrok 4.6—Not directly comparable
HealthBench Professional (raw)
Claude Opus 573.4%SourceGrok 4.6—Not directly comparable
BioMysteryBench (human-solvable)
Claude Opus 590.1%SourceGrok 4.6—Not directly comparable
BioMysteryBench (human-difficult)
Claude Opus 549.4%SourceGrok 4.6—Not directly comparable
SpatialBench Verified
Claude Opus 572.5%SourceGrok 4.6—Not directly comparable
SingleCellBench
Claude Opus 560.6%SourceGrok 4.6—Not directly comparable
ProteinGym Hard
Claude Opus 547.7%SourceGrok 4.6—Not directly comparable
Protein Design
Claude Opus 542.5%SourceGrok 4.6—Not directly comparable
Organic chemistry V2
Claude Opus 561.6%SourceGrok 4.6—Not directly comparable
Protocols (troubleshooting)
Claude Opus 561.1%SourceGrok 4.6—Not directly comparable
Protocols (understanding)
Claude Opus 578.4%SourceGrok 4.6—Not directly comparable
GPQA Diamond (Vals)
Claude Opus 593.4%SourceGrok 4.694.7%SourceGrok 4.6 leads this result
MMLU-Pro (Vals)
Claude Opus 591.6%SourceGrok 4.689.4%SourceClaude Opus 5 leads this result
Multilingual
Math
IMO 2026
Claude Opus 542/42SourceGrok 4.6—Not directly comparable
RiemannBench (no tools)
Claude Opus 560.0%SourceGrok 4.6—Not directly comparable
RiemannBench (tools)
Claude Opus 579.0%SourceGrok 4.6—Not directly comparable
ArXivMath Jun. 2026 (no tools)
Claude Opus 590.8%SourceGrok 4.6—Not directly comparable
ArXivMath Jun. 2026 (tools)
Claude Opus 591.3%SourceGrok 4.6—Not directly comparable
75 public results · 16 shared
Watch Claude Opus 5 vs Grok 4.6
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.
Last updated September 25, 2026