Skip to main content
BenchLM
Advertise

Model comparison

Inkling-Small vs Qwen3.5 397B

Updated July 30, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

Inkling-Small

Thinking Machines Lab

—

Evidence status unavailable

90% interval unavailable

Qwen3.5 397B

Alibaba

56.2/100

Estimated · Public rank #78

90% interval 44.7–67.7

At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.

11 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.

  • Agentic work

    Tool use, computer use, and multi-step task completion

    Inkling-Small

    Inkling-Small leads on the same 2 weighted benchmark rows.

    Confidence: limited

  • 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

    The category averages use different weighted benchmark sets, so they are 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. Qwen3.5 397B does not fit this workload in one request. Qwen3.5 397B 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.

Shared results
11
Inkling-Small only
9
Qwen3.5 397B only
27
Like-for-like categories
3 / 8

2 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.

Agentic

Like-for-like
Inkling-Small
70.1
Qwen3.5 397B
56.5
Weighted basis
2 vs 2 rows
Reading
Inkling-Small leads

Math

Like-for-like
Inkling-Small
92.9
Qwen3.5 397B
90.6
Weighted basis
2 vs 2 rows
Reading
Inkling-Small leads

Multimodal

Like-for-like
Inkling-Small
76.6
Qwen3.5 397B
79.6
Weighted basis
2 vs 2 rows
Reading
Qwen3.5 397B leads

Coding

Directional only
Inkling-Small
62.4
Qwen3.5 397B
66.5
Weighted basis
3 vs 2 rows
Reading
Directional only

Knowledge

Directional only
Inkling-Small
53.4
Qwen3.5 397B
56.6
Weighted basis
2 vs 4 rows
Reading
Directional only

Reasoning

Not comparable
Inkling-Small
40.1
Qwen3.5 397B
63.2
Weighted basis
1 vs 1 rows
Reading
Not comparable

Multilingual

Not comparable
Inkling-Small
Not measured
Qwen3.5 397B
84.7
Weighted basis
0 vs 1 rows
Reading
Not comparable

Instruction following

Not comparable
Inkling-Small
82.2
Qwen3.5 397B
92.6
Weighted basis
1 vs 1 rows
Reading
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.

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

Inkling-Small
$0.0013
Fits in one request
Qwen3.5 397B
$0.0024
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

Inkling-Small
$0.03332
Fits in one request
Qwen3.5 397B
$0.0408
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

Inkling-Small
$0.0492
Fits in one request
Qwen3.5 397B
$0.168
Does not fit in one request
Cached input priced at the published list-input rate

Qwen3.5 397B does not fit this workload in one request. Qwen3.5 397B 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.

Context window

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

Inkling-Small

1M

Qwen3.5 397B

128K

API model ID

Inkling-Small

Not sourced

Qwen3.5 397B

Not sourced

Cached-input rate

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

Inkling-Small

$0.116 per 1M cached input tokens

Qwen3.5 397B

Not published

Documented inputs

Inkling-Small

Not sourced

Qwen3.5 397B

Not sourced

Documented outputs

Inkling-Small

Not sourced

Qwen3.5 397B

Not sourced

Provider availability

Inkling-Small

Not sourced

Qwen3.5 397B

Not sourced

Reasoning profile

Inkling-Small

Hybrid

Qwen3.5 397B

Non-Reasoning

Weight access

Inkling-Small

Open Weight

Qwen3.5 397B

Open Weight

License

Inkling-Small

Open Weight

Qwen3.5 397B

Open Weight

Release date

Inkling-Small

2026-07-30

Qwen3.5 397B

2026-02-16

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: $0.03332 vs $0.0408. Cache-heavy agent loop: $0.0492 vs $0.168.
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 evidence47 rows

Agentic

  • Terminal-Bench 2.0

    Inkling-Small64.7%
    Source
    Qwen3.5 397B52.5%
    Source

    Inkling-Small leads this result

  • BrowseComp

    Inkling-Small77.4%
    Source
    Qwen3.5 397B62%
    Source

    Inkling-Small leads this result

  • MCP Atlas

    Inkling-Small79.6%
    Source
    Qwen3.5 397B46.1%
    Source

    Inkling-Small leads this result

  • Toolathlon-Verified

    Inkling-Small54.4%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • Claw-Eval

    Inkling-Small—
    Qwen3.5 397B56.8%
    Source

    Not directly comparable

  • QwenClawBench

    Inkling-Small—
    Qwen3.5 397B51.8%
    Source

    Not directly comparable

  • τ³-bench results

    Inkling-Small—
    Qwen3.5 397B68.4%
    Source

    Not directly comparable

  • VITA-Bench

    Inkling-Small—
    Qwen3.5 397B43.7%
    Source

    Not directly comparable

  • DeepPlanning

    Inkling-Small—
    Qwen3.5 397B37.6%
    Source

    Not directly comparable

  • Toolathlon

    Inkling-Small—
    Qwen3.5 397B36.3%
    Source

    Not directly comparable

  • MCP-Tasks

    Inkling-Small—
    Qwen3.5 397B74.2%
    Source

    Not directly comparable

  • WideResearch

    Inkling-Small—
    Qwen3.5 397B74.0%
    Source

    Not directly comparable

  • Gert Labs

    Inkling-Small—
    Qwen3.5 397B46.76%
    Source

    Not directly comparable

  • ResearchClawBench

    Inkling-Small—
    Qwen3.5 397B14.2%
    Source

    Not directly comparable

Coding

  • SWE-bench Verified

    Inkling-Small80.2%
    Source
    Qwen3.5 397B76.2%
    Source

    Inkling-Small leads this result

  • SWE-bench Pro

    Inkling-Small55.9%
    Source
    Qwen3.5 397B50.9%
    Source

    Inkling-Small leads this result

  • Terminal-Bench 2.0

    Inkling-Small64.7%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • SciCode

    Inkling-Small48.7%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • LiveCodeBench v6

    Inkling-Small—
    Qwen3.5 397B83.6%
    Source

    Not directly comparable

Reasoning

  • ARC-AGI-2

    Inkling-Small40.1%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • CritPt

    Inkling-Small8.3%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • LongBench v2

    Inkling-Small—
    Qwen3.5 397B63.2%
    Source

    Not directly comparable

  • AI-Needle

    Inkling-Small—
    Qwen3.5 397B68.7%
    Source

    Not directly comparable

Knowledge

  • GPQA

    Inkling-Small89.5%
    Source
    Qwen3.5 397B88.4%
    Source

    Inkling-Small leads this result

  • GPQA-D

    Inkling-Small89.5%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • HLE

    Inkling-Small47.8%
    Source
    Qwen3.5 397B28.7%
    Source

    Inkling-Small leads this result

  • HLE w/o tools

    Inkling-Small31.6%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • SuperGPQA

    Inkling-Small—
    Qwen3.5 397B70.4%
    Source

    Not directly comparable

  • MMLU-Pro

    Inkling-Small—
    Qwen3.5 397B87.8%
    Source

    Not directly comparable

  • MMLU-Redux

    Inkling-Small—
    Qwen3.5 397B94.9%
    Source

    Not directly comparable

  • C-Eval

    Inkling-Small—
    Qwen3.5 397B93%
    Source

    Not directly comparable

Math

  • AIME26

    Inkling-Small95.5%
    Source
    Qwen3.5 397B93.3%
    Source

    Inkling-Small leads this result

  • HMMT Feb 2026

    Inkling-Small90.2%
    Source
    Qwen3.5 397B87.9%
    Source

    Inkling-Small leads this result

  • HMMT Feb 2025

    Inkling-Small—
    Qwen3.5 397B94.8%
    Source

    Not directly comparable

  • HMMT Nov 2025

    Inkling-Small—
    Qwen3.5 397B92.7%
    Source

    Not directly comparable

  • MMAnswerBench

    Inkling-Small—
    Qwen3.5 397B80.9%
    Source

    Not directly comparable

Multilingual

  • MMLU-ProX

    Inkling-Small—
    Qwen3.5 397B84.7%
    Source

    Not directly comparable

  • NOVA-63

    Inkling-Small—
    Qwen3.5 397B59.1%
    Source

    Not directly comparable

Multimodal

  • MMMU-Pro

    Inkling-Small74%
    Source
    Qwen3.5 397B79%
    Source

    Qwen3.5 397B leads this result

  • CharXiv

    Inkling-Small81.3%
    Source
    Qwen3.5 397B80.8%
    Source

    Inkling-Small leads this result

  • CharXiv w/o tools

    Inkling-Small77.4%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • MathVision

    Inkling-Small—
    Qwen3.5 397B88.6%
    Source

    Not directly comparable

  • VideoMMMU

    Inkling-Small—
    Qwen3.5 397B84.7%
    Source

    Not directly comparable

  • ScreenSpot Pro

    Inkling-Small—
    Qwen3.5 397B65.6%
    Source

    Not directly comparable

  • V*

    Inkling-Small—
    Qwen3.5 397B95.8%
    Source

    Not directly comparable

Instruction following

  • IFBench

    Inkling-Small82.2%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • IFEval

    Inkling-Small—
    Qwen3.5 397B92.6%
    Source

    Not directly comparable

Frequently asked questions

Which is better, Inkling-Small or Qwen3.5 397B?

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, Inkling-Small or Qwen3.5 397B?

The current coding 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 is better for agentic tasks, Inkling-Small or Qwen3.5 397B?

Inkling-Small leads the like-for-like agentic tasks comparison across 2 shared weighted benchmark rows.

Which costs less, Inkling-Small or Qwen3.5 397B?

For the stated presets, chat costs $0.0013 on Inkling-Small and $0.0024 on Qwen3.5 397B; repository review costs $0.03332 and $0.0408; the cache-heavy agent loop costs $0.0492 and $0.168. Qwen3.5 397B does not fit this workload in one request. Qwen3.5 397B has no published cached-input rate, so cached tokens use its listed input rate.

Which has the larger context window, Inkling-Small or Qwen3.5 397B?

Inkling-Small has the larger documented context window: 1M, compared with 128K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated July 30, 2026

Watch Inkling-Small vs Qwen3.5 397B

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

Read a sample issue

Join 2,000+ readers.