Multimodal
Like-for-like- Claude Opus 4.5
- 23.5
- #46/48
- Inkling-Small
- 48.8
- #39/48
- Basis
- Provisional lane · 2 vs 2 weighted rows
- Reading
- Inkling-Small leads
Claude Opus 4.5 vs Inkling-Small
Updated September 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Inkling-Small has the higher public score estimate, 59.25 versus 58.41, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
11 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.
Long documents
Prompts that approach the documented context limit
Inkling-Small
Inkling-Small has the larger documented context window.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
Inkling-Small
Inkling-Small 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
Inkling-Small
Inkling-Small 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
Claude Opus 4.5 is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Claude Opus 4.5 is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Confidence: limited
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 Opus 4.5 does not fit this workload in one request. Claude Opus 4.5 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.
34 Claude Opus 4.5 only
11 Shared
14 Inkling-Small only
- Shared results
- 11
- Claude Opus 4.5 only
- 34
- Inkling-Small only
- 14
- Like-for-like categories
- 2 / 8
3 categories rest 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.
Instruction following
Like-for-like- Claude Opus 4.5
- 39.5
- #105/123
- Inkling-Small
- 89.6
- #25/123
- Basis
- Provisional lane · 1 vs 1 weighted rows
- Reading
- Inkling-Small leads
Agentic
Directional only- Claude Opus 4.5
- 43.2
- Estimated · #97/152
- Inkling-Small
- 36.9
- Supported · #128/152
- Basis
- BenchAlign lane · 15 vs 5 public rows
- Reading
- Directional only
Coding
Directional only- Claude Opus 4.5
- 56.6
- Estimated · #36/151
- Inkling-Small
- 43.8
- Supported · #97/151
- Basis
- BenchAlign lane · 5 vs 6 public rows
- Reading
- Directional only
Knowledge
Directional only- Claude Opus 4.5
- 53.9
- Estimated · #61/183
- Inkling-Small
- 56.5
- Supported · #46/183
- Basis
- BenchAlign lane · 6 vs 6 public rows
- Reading
- Directional only
Reasoning
Not comparable- Claude Opus 4.5
- 72.1
- Unranked · 4 rankable rows
- Inkling-Small
- 42.7
- Unranked · 3 rankable rows
- Basis
- Provisional lane · 1 vs 1 weighted rows
- Reading
- Not comparable
Math
Not comparable- Claude Opus 4.5
- 58.3
- #6/7
- Inkling-Small
- 76.9
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 4 vs 2 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Opus 4.5
- 82.9
- #2/12
- Inkling-Small
- Not ranked
- Basis
- Provisional lane · 1 vs 0 weighted rows
- Reading
- Not comparable
| Category | Claude Opus 4.5 | Inkling-Small | Basis | Reading |
|---|---|---|---|---|
| Multimodal | 23.5#46/48 | 48.8#39/48 | Like-for-likeProvisional lane · 2 vs 2 weighted rows | Inkling-Small leads |
| Instruction following | 39.5#105/123 | 89.6#25/123 | Like-for-likeProvisional lane · 1 vs 1 weighted rows | Inkling-Small leads |
| Agentic | 43.2Estimated · #97/152 | 36.9Supported · #128/152 | Directional onlyBenchAlign lane · 15 vs 5 public rows | Directional only |
| Coding | 56.6Estimated · #36/151 | 43.8Supported · #97/151 | Directional onlyBenchAlign lane · 5 vs 6 public rows | Directional only |
| Knowledge | 53.9Estimated · #61/183 | 56.5Supported · #46/183 | Directional onlyBenchAlign lane · 6 vs 6 public rows | Directional only |
| Reasoning | 72.1Unranked · 4 rankable rows | 42.7Unranked · 3 rankable rows | Not comparableProvisional lane · 1 vs 1 weighted rows | Not comparable |
| Math | 58.3#6/7 | 76.9Unranked · 2 rankable rows | Not comparableProvisional lane · 4 vs 2 weighted rows | Not comparable |
| Multilingual | 82.9#2/12 | Not ranked | Not comparableProvisional lane · 1 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.
IFBench
Instruction following
Claude Opus 4.5: 58%Inkling-Small: 82.2%Normalized gap 24.2Claude Opus 4.5 source Inkling-Small sourceHLE
Knowledge
Claude Opus 4.5: 30.8%Inkling-Small: 47.8%Normalized gap 17.0Claude Opus 4.5 source Inkling-Small sourceCharXiv
Multimodal
Claude Opus 4.5: 68.5%Inkling-Small: 81.3%Normalized gap 12.8Claude Opus 4.5 source Inkling-Small sourceTerminal-Bench 2.0
Agentic
Claude Opus 4.5: 59.3%Inkling-Small: 64.7%Normalized gap 5.4Claude Opus 4.5 source Inkling-Small sourceHMMT Feb 2026
Math
Claude Opus 4.5: 85.3%Inkling-Small: 90.2%Normalized gap 4.9Claude Opus 4.5 source Inkling-Small 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 4.5
- $0.0175
- Fits in one request
- Inkling-Small
- $0.0013
- Fits in one request
Inkling-Small has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Opus 4.5
- $0.325
- Fits in one request
- Inkling-Small
- $0.03332
- Fits in one request
Inkling-Small 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 4.5
- $1.35
- Does not fit in one request
- Cached input priced at the published list-input rate
- Inkling-Small
- $0.0492
- Fits in one request
Claude Opus 4.5 does not fit this workload in one request. Claude Opus 4.5 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 Opus 4.5Inkling-Small
Context window
Maximum documented context; output-token limits may be lower.
Claude Opus 4.5
200K
Inkling-Small
1M
API model ID
Claude Opus 4.5
Not sourced
Inkling-Small
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Opus 4.5
Not published
Inkling-Small
$0.116 per 1M cached input tokens
Documented inputs
Claude Opus 4.5
Not sourced
Inkling-Small
Not sourced
Documented outputs
Claude Opus 4.5
Not sourced
Inkling-Small
Not sourced
Provider availability
Claude Opus 4.5
Not sourced
Inkling-Small
Not sourced
Reasoning profile
Claude Opus 4.5
Non-Reasoning
Inkling-Small
Hybrid
Weight access
Claude Opus 4.5
Proprietary
Inkling-Small
Open Weight
License
Claude Opus 4.5
Proprietary
Inkling-Small
Open Weight
Release date
Claude Opus 4.5
2025-11-01
Inkling-Small
2026-07-30
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
- Inkling-Small has the higher public score estimate, 59.25 versus 58.41, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.325 vs $0.03332. Cache-heavy agent loop: $1.35 vs $0.0492.
- Context tradeoff
- Inkling-Small 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 evidence59 rows
Agentic
Terminal-Bench 2.0
Claude Opus 4.559.3%SourceInkling-Small64.7%SourceInkling-Small leads this result
OSWorld-Verified
Claude Opus 4.566.3%SourceInkling-Small—Not directly comparable
OSWorld
Claude Opus 4.566.3%SourceInkling-Small—Not directly comparable
Claw-Eval
Claude Opus 4.559.6%SourceInkling-Small—Not directly comparable
QwenClawBench
Claude Opus 4.552.3%SourceInkling-Small—Not directly comparable
τ³-bench results
Claude Opus 4.570.2%SourceInkling-Small—Not directly comparable
VITA-Bench
Claude Opus 4.523.3%SourceInkling-Small—Not directly comparable
DeepPlanning
Claude Opus 4.526.4%SourceInkling-Small—Not directly comparable
Toolathlon
Claude Opus 4.543.5%SourceInkling-Small—Not directly comparable
MCP Atlas
Claude Opus 4.542.3%SourceInkling-Small79.6%SourceInkling-Small leads this result
MCP-Tasks
Claude Opus 4.571.8%SourceInkling-Small—Not directly comparable
WideResearch
Claude Opus 4.576.4%SourceInkling-Small—Not directly comparable
CyberGym
Claude Opus 4.550.6%SourceInkling-Small—Not directly comparable
Gert Labs
Claude Opus 4.564.23%SourceInkling-Small—Not directly comparable
JobBench
Claude Opus 4.532.3%SourceInkling-Small—Not directly comparable
BrowseComp
Claude Opus 4.5—Inkling-Small77.4%SourceNot directly comparable
Toolathlon-Verified
Claude Opus 4.5—Inkling-Small54.4%SourceNot directly comparable
Terminal-Bench 2.1 (Vals)
Claude Opus 4.5—Inkling-Small55.1%SourceNot directly comparable
Coding
SWE-bench Verified
Claude Opus 4.580.9%SourceInkling-Small80.2%SourceClaude Opus 4.5 leads this result
LiveCodeBench v6
Claude Opus 4.584.8%SourceInkling-Small—Not directly comparable
SWE-bench Pro
Claude Opus 4.557.1%SourceInkling-Small55.9%SourceClaude Opus 4.5 leads this result
SWE Multilingual
Claude Opus 4.577.5%SourceInkling-Small—Not directly comparable
NL2Repo
Claude Opus 4.543.2%SourceInkling-Small—Not directly comparable
Terminal-Bench 2.0
Claude Opus 4.5—Inkling-Small64.7%SourceNot directly comparable
SciCode
Claude Opus 4.5—Inkling-Small48.7%SourceNot directly comparable
LiveCodeBench (Vals)
Claude Opus 4.5—Inkling-Small85.9%SourceNot directly comparable
SWE-bench (Vals)
Claude Opus 4.5—Inkling-Small82.2%SourceNot directly comparable
Reasoning
Knowledge
GPQA
Claude Opus 4.587%SourceInkling-Small89.5%SourceInkling-Small leads this result
SuperGPQA
Claude Opus 4.570.6%SourceInkling-Small—Not directly comparable
MMLU-Pro
Claude Opus 4.589.5%SourceInkling-Small—Not directly comparable
MMLU-Redux
Claude Opus 4.596.6%SourceInkling-Small—Not directly comparable
C-Eval
Claude Opus 4.592.2%SourceInkling-Small—Not directly comparable
HLE
Claude Opus 4.530.8%SourceInkling-Small47.8%SourceInkling-Small leads this result
GPQA-D
Claude Opus 4.5—Inkling-Small89.5%SourceNot directly comparable
HLE w/o tools
Claude Opus 4.5—Inkling-Small31.6%SourceNot directly comparable
GPQA Diamond (Vals)
Claude Opus 4.5—Inkling-Small83.6%SourceNot directly comparable
MMLU-Pro (Vals)
Claude Opus 4.5—Inkling-Small85.6%SourceNot directly comparable
Math
AIME26
Claude Opus 4.595.1%SourceInkling-Small95.5%SourceInkling-Small leads this result
HMMT Feb 2025
Claude Opus 4.592.9%SourceInkling-Small—Not directly comparable
HMMT Nov 2025
Claude Opus 4.593.3%SourceInkling-Small—Not directly comparable
HMMT Feb 2026
Claude Opus 4.585.3%SourceInkling-Small90.2%SourceInkling-Small leads this result
MMAnswerBench
Claude Opus 4.584.0%SourceInkling-Small—Not directly comparable
FrontierMath v2 (Tiers 1-3)
Claude Opus 4.520.690%SourceInkling-Small—Not directly comparable
FrontierMath v2 (Tier 4)
Claude Opus 4.54.167%SourceInkling-Small—Not directly comparable
Multilingual
Multimodal
MMMU-Pro
Claude Opus 4.570.6%SourceInkling-Small74%SourceInkling-Small leads this result
MathVision
Claude Opus 4.574.3%SourceInkling-Small—Not directly comparable
CharXiv
Claude Opus 4.568.5%SourceInkling-Small81.3%SourceInkling-Small leads this result
VideoMMMU
Claude Opus 4.584.4%SourceInkling-Small—Not directly comparable
ScreenSpot Pro
Claude Opus 4.545.7%SourceInkling-Small—Not directly comparable
V*
Claude Opus 4.567.0%SourceInkling-Small—Not directly comparable
CharXiv w/o tools
Claude Opus 4.5—Inkling-Small77.4%SourceNot directly comparable
Instruction following
Frequently asked questions
Which is better, Claude Opus 4.5 or Inkling-Small?
Inkling-Small has the higher public score estimate, 59.25 versus 58.41, 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 Opus 4.5 or Inkling-Small?
Claude Opus 4.5 scores higher for coding on the public lane, 56.6 to 43.8. Claude Opus 4.5 is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which is better for agentic tasks, Claude Opus 4.5 or Inkling-Small?
Claude Opus 4.5 scores higher for agentic tasks on the public lane, 43.2 to 36.9. Claude Opus 4.5 is scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which costs less, Claude Opus 4.5 or Inkling-Small?
For the stated presets, chat costs $0.0175 on Claude Opus 4.5 and $0.0013 on Inkling-Small; repository review costs $0.325 and $0.03332; the cache-heavy agent loop costs $1.35 and $0.0492. Claude Opus 4.5 does not fit this workload in one request. Claude Opus 4.5 has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, Claude Opus 4.5 or Inkling-Small?
Inkling-Small has the larger documented context window: 1M, compared with 200K.
Related comparisons
Last updated September 10, 2026
Watch Claude Opus 4.5 vs Inkling-Small
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.