Agentic
Like-for-like- Claude Opus 5
- 77.4
- Supported · #2/151
- Gemini 3.1 Pro
- 40.1
- Supported · #120/151
- Basis
- BenchAlign lane · 19 vs 6 public rows
- Reading
- Claude Opus 5 leads
Claude Opus 5 vs Gemini 3.1 Pro
Updated September 4, 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, 80.66 versus 69.82, and the 90% score intervals do not overlap.
9 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Share or export
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 public coding lane, 75.6 to 46.3, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: stronger
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 40.1, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: stronger
Chat turn cost
1K fresh input + 500 output tokens
Gemini 3.1 Pro
Gemini 3.1 Pro 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
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Gemini 3.1 Pro
Gemini 3.1 Pro has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
Gemini 3.1 Pro
Gemini 3.1 Pro has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Long documents
Prompts that approach the documented context limit
No clear pick
The documented context windows are equal.
Confidence: documented
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
64 Claude Opus 5 only
9 Shared
19 Gemini 3.1 Pro only
- Shared results
- 9
- Claude Opus 5 only
- 64
- Gemini 3.1 Pro only
- 19
- Like-for-like categories
- 3 / 8
1 category rests on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign 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.
Coding
Like-for-like- Claude Opus 5
- 75.6
- Supported · #3/183
- Gemini 3.1 Pro
- 46.3
- Supported · #99/183
- Basis
- BenchAlign lane · 16 vs 5 public rows
- Reading
- Claude Opus 5 leads
Knowledge
Like-for-like- Claude Opus 5
- 82.1
- Supported · #3/181
- Gemini 3.1 Pro
- 65.6
- Supported · #24/181
- Basis
- BenchAlign lane · 19 vs 6 public rows
- Reading
- Claude Opus 5 leads · intervals overlap
Multimodal
Directional only- Claude Opus 5
- 88.7
- #2/48
- Gemini 3.1 Pro
- 79.2
- #12/48
- Basis
- Provisional lane · 1 vs 2 weighted rows
- Reading
- Directional only
Reasoning
Not comparable- Claude Opus 5
- 75.6
- #13/22
- Gemini 3.1 Pro
- 50.7
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 2 vs 2 weighted rows
- Reading
- Not comparable
Math
Not comparable- Claude Opus 5
- Not ranked
- Gemini 3.1 Pro
- 54.3
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Opus 5
- Not ranked
- Gemini 3.1 Pro
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Opus 5
- Not ranked
- Gemini 3.1 Pro
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | Claude Opus 5 | Gemini 3.1 Pro | Basis | Reading |
|---|---|---|---|---|
| Agentic | 77.4Supported · #2/151 | 40.1Supported · #120/151 | Like-for-likeBenchAlign lane · 19 vs 6 public rows | Claude Opus 5 leads |
| Coding | 75.6Supported · #3/183 | 46.3Supported · #99/183 | Like-for-likeBenchAlign lane · 16 vs 5 public rows | Claude Opus 5 leads |
| Knowledge | 82.1Supported · #3/181 | 65.6Supported · #24/181 | Like-for-likeBenchAlign lane · 19 vs 6 public rows | Claude Opus 5 leads · intervals overlap |
| Multimodal | 88.7#2/48 | 79.2#12/48 | Directional onlyProvisional lane · 1 vs 2 weighted rows | Directional only |
| Reasoning | 75.6#13/22 | 50.7Unranked · 2 rankable rows | Not comparableProvisional lane · 2 vs 2 weighted rows | Not comparable |
| Math | Not ranked | 54.3Unranked · 2 rankable rows | Not comparableProvisional lane · 0 vs 2 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 |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
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-3
Reasoning
Claude Opus 5: 30.2%Gemini 3.1 Pro: 0.4%Normalized gap 29.7Claude Opus 5 source Gemini 3.1 Pro sourceARC-AGI-2
Reasoning
Claude Opus 5: 90.4%Gemini 3.1 Pro: 77.1%Normalized gap 13.3Claude Opus 5 source Gemini 3.1 Pro sourceHLE w/o tools
Knowledge
Claude Opus 5: 56.3%Gemini 3.1 Pro: 45.4%Normalized gap 10.9Claude Opus 5 source Gemini 3.1 Pro sourceMMLU-Pro (Vals)
Knowledge
Claude Opus 5: 91.6%Gemini 3.1 Pro: 91.0%Normalized gap 0.6Claude Opus 5 source Gemini 3.1 Pro sourceLiveCodeBench (Vals)
Coding
Claude Opus 5: 89.0%Gemini 3.1 Pro: 88.5%Normalized gap 0.5Claude Opus 5 source Gemini 3.1 Pro 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 5
- $0.0175
- Fits in one request
- Gemini 3.1 Pro
- $0.008
- Fits in one request
Gemini 3.1 Pro 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
- Gemini 3.1 Pro
- $0.136
- Fits in one request
Gemini 3.1 Pro 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
- Gemini 3.1 Pro
- $0.2
- Fits in one request
Gemini 3.1 Pro has the lower modeled cost
Costs use the listed standard API rates.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationClaude Opus 5Gemini 3.1 Pro
Context window
Maximum documented context; output-token limits may be lower.
Claude Opus 5
Gemini 3.1 Pro
API model ID
Claude Opus 5
claude-opus-5
Anthropic model overviewGemini 3.1 Pro
gemini-3.1-pro-preview
Google Gemini 3.1 Pro Preview model documentationCached-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 pricingGemini 3.1 Pro
$0.2 per 1M cached input tokens
Google Gemini API pricingDocumented inputs
Claude Opus 5
text, image
Anthropic model overviewGemini 3.1 Pro
text, image, video, audio, pdf
Google Gemini 3.1 Pro Preview model documentationDocumented outputs
Claude Opus 5
Gemini 3.1 Pro
Provider availability
Claude Opus 5
Generally Available · Claude API
Anthropic model overviewGemini 3.1 Pro
Preview · Gemini API, Google AI Studio
Google Gemini model catalogReasoning profile
Claude Opus 5
Reasoning
Gemini 3.1 Pro
Reasoning
Weight access
Claude Opus 5
Proprietary
Gemini 3.1 Pro
Proprietary
License
Claude Opus 5
Proprietary
Gemini 3.1 Pro
Proprietary
Release date
Claude Opus 5
2026-07-24
Gemini 3.1 Pro
2026-02-19
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, 80.66 versus 69.82, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.325 vs $0.136. Cache-heavy agent loop: $0.45 vs $0.2.
- 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 evidence92 rows
Agentic
Terminal-Bench 3.0
Claude Opus 542.7%SourceGemini 3.1 Pro—Not directly comparable
BrowseComp
Claude Opus 590.8%SourceGemini 3.1 Pro—Not directly comparable
HLE w/ tools
Claude Opus 564.7%SourceGemini 3.1 Pro—Not directly comparable
DeepSearchQA
Claude Opus 595.0%SourceGemini 3.1 Pro69.7%SourceClaude Opus 5 leads this result
DRACO
Claude Opus 588.6%SourceGemini 3.1 Pro—Not directly comparable
BrowseComp (10-agent, prerelease)
Claude Opus 593.6%SourceGemini 3.1 Pro—Not directly comparable
OSWorld 2.0
Claude Opus 570.6%SourceGemini 3.1 Pro—Not directly comparable
MCP Atlas
Claude Opus 585.8%SourceGemini 3.1 Pro—Not directly comparable
MCP-Atlas claim coverage
Claude Opus 589.1%SourceGemini 3.1 Pro—Not directly comparable
LAB all-pass (Anthropic harness)
Claude Opus 523.58%SourceGemini 3.1 Pro—Not directly comparable
LAB criterion-pass (Anthropic harness)
Claude Opus 593.74%SourceGemini 3.1 Pro—Not directly comparable
LAB all-pass (Harvey held-out)
Claude Opus 511.7%SourceGemini 3.1 Pro—Not directly comparable
LAB criterion-pass (Harvey held-out)
Claude Opus 594.1%SourceGemini 3.1 Pro—Not directly comparable
Toolathlon-Verified
Claude Opus 580.6%SourceGemini 3.1 Pro—Not directly comparable
Toolathlon Verified Pass@3
Claude Opus 587.0%SourceGemini 3.1 Pro—Not directly comparable
Toolathlon Verified Pass³
Claude Opus 573.1%SourceGemini 3.1 Pro—Not directly comparable
Toolathlon Verified avg. turns
Claude Opus 523.5 turnsSourceGemini 3.1 Pro—Not directly comparable
AutomationBench
Claude Opus 526.0%SourceGemini 3.1 Pro—Not directly comparable
Terminal-Bench 2.1 (Vals)
Claude Opus 584.6%SourceGemini 3.1 Pro70.8%SourceClaude Opus 5 leads this result
Claw-Eval
Claude Opus 5—Gemini 3.1 Pro57.8%SourceNot directly comparable
τ²-bench results
Claude Opus 5—Gemini 3.1 Pro95.6%SourceNot directly comparable
Gert Labs
Claude Opus 5—Gemini 3.1 Pro56.87%SourceNot directly comparable
ResearchClawBench
Claude Opus 5—Gemini 3.1 Pro13.3%SourceNot directly comparable
Coding
Bug Hunt Bench
Claude Opus 527 fixesSourceGemini 3.1 Pro—Not directly comparable
SWE-bench Verified
Claude Opus 596%SourceGemini 3.1 Pro—Not directly comparable
SWE-bench Pro
Claude Opus 579.2%SourceGemini 3.1 Pro—Not directly comparable
SWE Multilingual
Claude Opus 589.5%SourceGemini 3.1 Pro—Not directly comparable
SWE Multimodal
Claude Opus 559.4%SourceGemini 3.1 Pro—Not directly comparable
deepSwe
Claude Opus 568.8%SourceGemini 3.1 Pro—Not directly comparable
FrontierCode 1.1 Main
Claude Opus 553.4%SourceGemini 3.1 Pro—Not directly comparable
FrontierCode 1.1 Extended
Claude Opus 563.6%SourceGemini 3.1 Pro—Not directly comparable
FrontierSWE v2
Claude Opus 552.0%SourceGemini 3.1 Pro—Not directly comparable
ProgramBench (episode 1)
Claude Opus 583.0%SourceGemini 3.1 Pro—Not directly comparable
ProgramBench
Claude Opus 593.0%SourceGemini 3.1 Pro—Not directly comparable
cursorBench32
Claude Opus 570.0%SourceGemini 3.1 Pro—Not directly comparable
VulcanBench v3
Claude Opus 587.0%SourceGemini 3.1 Pro—Not directly comparable
VulcanBench CII v1
Claude Opus 596.4%SourceGemini 3.1 Pro—Not directly comparable
LiveCodeBench (Vals)
Claude Opus 589.0%SourceGemini 3.1 Pro88.5%SourceClaude Opus 5 leads this result
SWE-bench (Vals)
Claude Opus 597.0%SourceGemini 3.1 Pro78.8%SourceClaude Opus 5 leads this result
LiveCodeBench Pro
Claude Opus 5—Gemini 3.1 Pro82.9%SourceNot directly comparable
React Native Evals
Claude Opus 5—Gemini 3.1 Pro78.9%SourceNot directly comparable
Vibe Code Bench
Claude Opus 5—Gemini 3.1 Pro32.03%SourceNot directly comparable
Reasoning
Knowledge
HLE
Claude Opus 564.7%SourceGemini 3.1 Pro—Not directly comparable
HLE w/o tools
Claude Opus 556.3%SourceGemini 3.1 Pro45.4%SourceClaude Opus 5 leads this result
HLE-Verified
Claude Opus 554.4%SourceGemini 3.1 Pro—Not directly comparable
LABBench2
Claude Opus 584.2%SourceGemini 3.1 Pro—Not directly comparable
HealthBench (raw)
Claude Opus 567.1%SourceGemini 3.1 Pro—Not directly comparable
HealthBench (length-adjusted)
Claude Opus 557.8%SourceGemini 3.1 Pro—Not directly comparable
HealthBench Professional
Claude Opus 559.8%SourceGemini 3.1 Pro—Not directly comparable
HealthBench Professional (raw)
Claude Opus 573.4%SourceGemini 3.1 Pro—Not directly comparable
BioMysteryBench (human-solvable)
Claude Opus 590.1%SourceGemini 3.1 Pro—Not directly comparable
BioMysteryBench (human-difficult)
Claude Opus 549.4%SourceGemini 3.1 Pro—Not directly comparable
SpatialBench Verified
Claude Opus 572.5%SourceGemini 3.1 Pro—Not directly comparable
SingleCellBench
Claude Opus 560.6%SourceGemini 3.1 Pro—Not directly comparable
ProteinGym Hard
Claude Opus 547.7%SourceGemini 3.1 Pro—Not directly comparable
Protein Design
Claude Opus 542.5%SourceGemini 3.1 Pro—Not directly comparable
Organic chemistry V2
Claude Opus 561.6%SourceGemini 3.1 Pro—Not directly comparable
Protocols (troubleshooting)
Claude Opus 561.1%SourceGemini 3.1 Pro—Not directly comparable
Protocols (understanding)
Claude Opus 578.4%SourceGemini 3.1 Pro—Not directly comparable
GPQA Diamond (Vals)
Claude Opus 593.4%SourceGemini 3.1 Pro95.5%SourceGemini 3.1 Pro leads this result
MMLU-Pro (Vals)
Claude Opus 591.6%SourceGemini 3.1 Pro91.0%SourceClaude Opus 5 leads this result
GPQA-D
Claude Opus 5—Gemini 3.1 Pro94.3%SourceNot directly comparable
HealthBench Hard
Claude Opus 5—Gemini 3.1 Pro20.6%SourceNot directly comparable
MedXpertQA (Text)
Claude Opus 5—Gemini 3.1 Pro71.5%SourceNot directly comparable
Math
IMO 2026
Claude Opus 542/42SourceGemini 3.1 Pro—Not directly comparable
RiemannBench (no tools)
Claude Opus 560.0%SourceGemini 3.1 Pro—Not directly comparable
RiemannBench (tools)
Claude Opus 579.0%SourceGemini 3.1 Pro—Not directly comparable
ArXivMath Jun. 2026 (no tools)
Claude Opus 590.8%SourceGemini 3.1 Pro—Not directly comparable
ArXivMath Jun. 2026 (tools)
Claude Opus 591.3%SourceGemini 3.1 Pro—Not directly comparable
FrontierMath v2 (Tiers 1-3)
Claude Opus 5—Gemini 3.1 Pro36.900%SourceNot directly comparable
FrontierMath v2 (Tier 4)
Claude Opus 5—Gemini 3.1 Pro16.700%SourceNot directly comparable
Multilingual
Multimodal
Chartography (no tools)
Claude Opus 529.6%SourceGemini 3.1 Pro—Not directly comparable
Chartography (tools)
Claude Opus 583.0%SourceGemini 3.1 Pro—Not directly comparable
BenchCAD Vision2Code (no tools)
Claude Opus 50.366SourceGemini 3.1 Pro—Not directly comparable
BenchCAD Vision2Code (tools)
Claude Opus 50.821SourceGemini 3.1 Pro—Not directly comparable
GDP.pdf (no tools)
Claude Opus 583.4%SourceGemini 3.1 Pro—Not directly comparable
GDP.pdf (tools)
Claude Opus 585.5%SourceGemini 3.1 Pro—Not directly comparable
OfficeQA
Claude Opus 578.1%SourceGemini 3.1 Pro—Not directly comparable
OfficeQA Pro
Claude Opus 566.9%SourceGemini 3.1 Pro—Not directly comparable
MMMU-Pro
Claude Opus 5—Gemini 3.1 Pro83.9%SourceNot directly comparable
CharXiv
Claude Opus 5—Gemini 3.1 Pro80.2%SourceNot directly comparable
ERQA
Claude Opus 5—Gemini 3.1 Pro69.4%SourceNot directly comparable
SimpleVQA
Claude Opus 5—Gemini 3.1 Pro72.4%SourceNot directly comparable
ScreenSpot Pro
Claude Opus 5—Gemini 3.1 Pro84.4%SourceNot directly comparable
ZeroBench
Claude Opus 5—Gemini 3.1 Pro29.0%SourceNot directly comparable
MedXpertQA (MM)
Claude Opus 5—Gemini 3.1 Pro81.3%SourceNot directly comparable
Frequently asked questions
Which is better, Claude Opus 5 or Gemini 3.1 Pro?
Claude Opus 5 has the higher public score, 80.66 versus 69.82, 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 Gemini 3.1 Pro?
Claude Opus 5 leads the public coding lane, 75.6 to 46.3, with Supported evidence for both models and non-overlapping 90% intervals.
Which is better for agentic tasks, Claude Opus 5 or Gemini 3.1 Pro?
Claude Opus 5 leads the public agentic tasks lane, 77.4 to 40.1, with Supported evidence for both models and non-overlapping 90% intervals.
Which costs less, Claude Opus 5 or Gemini 3.1 Pro?
For the stated presets, chat costs $0.0175 on Claude Opus 5 and $0.008 on Gemini 3.1 Pro; repository review costs $0.325 and $0.136; 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 Gemini 3.1 Pro?
Both models list the same context window, 1M.
Related comparisons
Last updated September 4, 2026
Watch Claude Opus 5 vs Gemini 3.1 Pro
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.