Hy4 vs Inkling

Updated November 4, 2026. We do not rank this pair: at least one has no public score. Public scores include evidence status and uncertainty.

Decision reading

At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. 2 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.

Model A

Hy4

Tencent

—

Evidence status unavailable

90% interval unavailable

Model B

Inkling

Thinking Machines Lab

47.91/100

Supported · Public rank #108

90% interval 29.8–66.0

Shared results: 2
Hy4 only: 13
Inkling only: 22
Like-for-like categories: 0 / 8

Supported: InklingHow 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.

Chat turn cost
1K fresh input + 500 output tokens
Hy4
Hy4 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Hy4
Hy4 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
Hy4
Hy4 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
Hy4 is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Hy4 is not ranked on the public lane for agentic, so no winner is named for agentic.
Confidence: limited
Long documents
Prompts that approach the documented context limit
No clear pick
The documented context windows are equal.
Confidence: documented

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.

Pick the job you would use them for

The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.

—Hy433.0Inkling

Not comparable · BenchAlign v5.8

The coding row is not comparable on the public lane: at least one model is not measured or not ranked there.

Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.

Open the coding leaderboard Shortlist models for this job

Same basis rules as the category table below

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.

HLEKnowledge
Normalized gap 13.1
Hy4:59.1%
Inkling:46%
Hy4 source Inkling source
MMMU-ProMultimodal
Normalized gap 5.1
Hy4:78.6%
Inkling:73.5%
Hy4 source Inkling source

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 v5.8 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

Not comparable

Hy4: Not ranked
Inkling: 37.0; Supported · #81/135
Basis: BenchAlign v5.8 lane · 7 vs 5 public rows
Reading: Not comparable

Coding

Not comparable

Hy4: Not ranked
Inkling: 33.0; Supported · #90/156
Basis: BenchAlign v5.8 lane · 5 vs 6 public rows
Reading: Not comparable

Reasoning

Not comparable

Hy4: Not ranked
Inkling: 78.1; #11/26
Basis: Provisional lane · 0 vs 0 weighted rows
Reading: Not comparable

Multimodal

Not comparable

Hy4: Not ranked
Inkling: 68.0; #38/58
Basis: Provisional lane · 1 vs 2 weighted rows
Reading: Not comparable

Knowledge

Not comparable

Hy4: Not ranked
Inkling: 50.6; Supported · #76/179
Basis: BenchAlign v5.8 lane · 1 vs 6 public rows
Reading: Not comparable

Multilingual

Not comparable

Hy4: Not ranked
Inkling: Not ranked
Basis: Provisional lane · 0 vs 0 weighted rows
Reading: Not comparable

Instruction following

Not comparable

Hy4: Not ranked
Inkling: 87.6; #40/157
Basis: Provisional lane · 0 vs 1 weighted rows
Reading: Not comparable

Math

Not comparable

Hy4: Not ranked
Inkling: 58.3; #11/12
Basis: Provisional lane · 0 vs 3 weighted rows
Reading: Not comparable

Category scores with the server-provided evidence basis for Hy4 and Inkling
Category	Hy4	Inkling	Basis	Reading
Agentic	Not ranked	37.0Supported · #81/135	Not comparableBenchAlign v5.8 lane · 7 vs 5 public rows	Not comparable
Coding	Not ranked	33.0Supported · #90/156	Not comparableBenchAlign v5.8 lane · 5 vs 6 public rows	Not comparable
Reasoning	Not ranked	78.1#11/26	Not comparableProvisional lane · 0 vs 0 weighted rows	Not comparable
Multimodal	Not ranked	68.0#38/58	Not comparableProvisional lane · 1 vs 2 weighted rows	Not comparable
Knowledge	Not ranked	50.6Supported · #76/179	Not comparableBenchAlign v5.8 lane · 1 vs 6 public rows	Not comparable
Multilingual	Not ranked	Not ranked	Not comparableProvisional lane · 0 vs 0 weighted rows	Not comparable
Instruction following	Not ranked	87.6#40/157	Not comparableProvisional lane · 0 vs 1 weighted rows	Not comparable
Math	Not ranked	58.3#11/12	Not comparableProvisional lane · 0 vs 3 weighted rows	Not comparable

Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.8) 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

Hy4: $0.00208; Fits in one request
Inkling: $0.00421; Fits in one request

Hy4 has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

Hy4: $0.0492; Fits in one request
Inkling: $0.10754; Fits in one request

Hy4 has the lower modeled cost

Costs use the listed standard API rates.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

Hy4: $0.05009; Fits in one request
Inkling: $0.159; Fits in one request

Hy4 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.

SpecificationHy4Inkling

Context window

Maximum documented context; output-token limits may be lower.

Hy4

Tencent Cloud TokenHub model pricing

Inkling

API model ID

Hy4

tencent/hy4

Tencent Cloud TokenHub model pricing

Inkling

Not sourced

Cached-input rate

A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.

Hy4

$0.042 per 1M cached input tokens

Tencent Cloud TokenHub model pricing

Inkling

$0.374 per 1M cached input tokens

Documented inputs

Hy4

Not sourced

Inkling

Not sourced

Documented outputs

Hy4

Not sourced

Inkling

Not sourced

Provider availability

Hy4

Not sourced

Inkling

Not sourced

Reasoning profile

Hy4

Reasoning

Inkling

Hybrid

Weight access

Hy4

Pending

Inkling

Open Weight

License

Hy4

Pending

Inkling

Open Weight

Release date

Hy4

2026-11-04

Inkling

2026-07-15

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: At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
Workload cost: Repository review: $0.0492 vs $0.10754. Cache-heavy agent loop: $0.05009 vs $0.159.
Context tradeoff: Both models list 1M.

Run the same representative tasks against both endpoints before changing production traffic.

Questions

Which is better, Hy4 or Inkling?

At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. The page therefore keeps the decision tied to the specific documented workload.

Which is better for coding, Hy4 or Inkling?

Hy4 is not ranked on the public lane for coding, so no winner is named for coding.

Which is better for agentic tasks, Hy4 or Inkling?

Hy4 is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.

Which costs less, Hy4 or Inkling?

For the stated presets, chat costs $0.00209 on Hy4 and $0.00421 on Inkling; repository review costs $0.0492 and $0.10754; the cache-heavy agent loop costs $0.05009 and $0.159. Costs use the listed standard API rates.

Which has the larger context window, Hy4 or Inkling?

Both models list the same context window, 1M.

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Hy4

API / mo$2,501

Self-host / moNot listed

Break-even—

Proprietary model — self-hosting not applicable.

Inkling

API / mo$4,913

Self-host / mo$27,962

Break-even344M/day

Model the full break-even

Benchmark evidence

The full public result ledger is available for audit without forcing a wide desktop table onto a phone.

Browse raw public benchmark evidence37 rows

Agentic

Terminal-Bench 4.0
Hy444.40%
Source
Inkling—
Not directly comparable
Terminal-Bench-Science 0.1
Hy419.8%
Source
Inkling—
Not directly comparable
HLE w/ tools
Hy459.1%
Source
Inkling—
Not directly comparable
Agents' Last Exam
Hy441.0%
Source
Inkling—
Not directly comparable
APEX-Agents
Hy475.2%
Source
Inkling—
Not directly comparable
AutomationBench
Hy456.2%
Source
Inkling—
Not directly comparable
OSWorld 2.1
Hy478.9%
Source
Inkling—
Not directly comparable
Terminal-Bench 2.1
Hy4—
Inkling63.8%
Source
Not directly comparable
BrowseComp
Hy4—
Inkling77.1%
Source
Not directly comparable
MCP Atlas
Hy4—
Inkling74.1%
Source
Not directly comparable
Terminal-Bench 2.1 (Vals)
Hy4—
Inkling47.6%
Source
Not directly comparable
CWE-bench v1
Hy4—
Inkling37.0%
Source
Not directly comparable

Coding

DeepSWE
Hy474.9%
Source
Inkling—
Not directly comparable
sweMarathon
Hy443.8%
Source
Inkling—
Not directly comparable
ProgramBench
Hy430.0%
Source
Inkling—
Not directly comparable
FrontierCode 2.0
Hy431.0%
Source
Inkling—
Not directly comparable
PostTrainBench v1.2
Hy435.2%
Source
Inkling—
Not directly comparable
SWE-bench Verified
Hy4—
Inkling77.6%
Source
Not directly comparable
SWE-bench Pro
Hy4—
Inkling54.3%
Source
Not directly comparable
Terminal-Bench 2.1
Hy4—
Inkling63.8%
Source
Not directly comparable
FrontierSWE v2
Hy4—
Inkling4.1%
Source
Not directly comparable
LiveCodeBench (Vals)
Hy4—
Inkling85.5%
Source
Not directly comparable
SWE-bench (Vals)
Hy4—
Inkling77.6%
Source
Not directly comparable

Multimodal

MMMU-Pro
Hy478.6%
Source
Inkling73.5%
Source
Hy4 leads this result
ZeroBench w/ Python
Hy455.4%
Source
Inkling—
Not directly comparable
CharXiv
Hy4—
Inkling82%
Source
Not directly comparable
CharXiv w/o tools
Hy4—
Inkling78.1%
Source
Not directly comparable

Knowledge

HLE
Hy459.1%
Source
Inkling46%
Source
Hy4 leads this result
GPQA
Hy4—
Inkling87.9%
Source
Not directly comparable
GPQA-D
Hy4—
Inkling87.9%
Source
Not directly comparable
HLE w/o tools
Hy4—
Inkling30%
Source
Not directly comparable
GPQA Diamond (Vals)
Hy4—
Inkling87.1%
Source
Not directly comparable
MMLU-Pro (Vals)
Hy4—
Inkling86.3%
Source
Not directly comparable

Instruction following

IFBench
Hy4—
Inkling79.8%
Source
Not directly comparable

Math

AIME26
Hy4—
Inkling97.1%
Source
Not directly comparable
FrontierMath v2 (Tiers 1-3)
Hy4—
Inkling33.333%
Source
Not directly comparable
FrontierMath v2 (Tier 4)
Hy4—
Inkling4.878%
Source
Not directly comparable

37 public results · 2 shared

Related comparisons

Compare API pricing Read the methodology Open the model selector

Watch Hy4 vs Inkling

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 6,100+ subscribers.

Last updated November 4, 2026

Hy4 vs Inkling

Decision reading

Which one for your work

Chat turn cost

Cache-heavy agent loop cost

Repository review cost

Coding work

Agentic work

Long documents

Which one for a specific job

What is actually comparable

Shape of the matched evidence

Category results, on a stated basis

Agentic

Coding

Reasoning

Multimodal

Knowledge

Multilingual

Instruction following

Math

What each workload costs

Chat turn

Repository review

Cache-heavy agent loop

Specification differences

Context window

API model ID

Cached-input rate

Documented inputs

Documented outputs

Provider availability

Reasoning profile

Weight access

License

Release date

If you already use one of these models

Questions

Which is better, Hy4 or Inkling?

Which is better for coding, Hy4 or Inkling?

Which is better for agentic tasks, Hy4 or Inkling?

Which costs less, Hy4 or Inkling?

Which has the larger context window, Hy4 or Inkling?

Self-host vs API cost

Benchmark evidence

Agentic

Coding

Multimodal

Knowledge

Instruction following

Math

Related comparisons

Watch Hy4 vs Inkling