Skip to main content
BenchLM
Data

Hy4 preview vs Mistral Large 4

Updated November 21, 2026. Rank says Hy4 preview is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.

Share or export
Share on XLinkedInAPI/MCP

Decision reading

Hy4 preview has the higher public point estimate, 68.2 versus 56.9. Their conditional score ranges overlap. These ranges do not establish rank confidence. 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
Tencent logo

Tencent

68.2/100

Supported · Public rank #26

Conditional range 61.8–74.7

Model B
Mistral logo

Mistral

56.9/100

Supported · Public rank #67

Conditional range 51.2–62.5

API price per 1M tokens

Hy4 preview: No comparable API price

Mistral Large 4: $0.70 input / $2.10 output

Context window

Hy4 preview: 1M

Mistral Large 4: 1M

Released

Hy4 preview: August 28, 2026

Mistral Large 4: October 6, 2026

Shared results
2
Hy4 preview only
32
Mistral Large 4 only
9
Like-for-like categories
2 / 8
Supported: Hy4 preview and Mistral Large 4. Conditional ranges do not establish rank confidence.How 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

    Hy4 preview

    Hy4 preview has the higher public agentic point estimate, 66.8 to 56.2, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

    Confidence: limited
Show secondary and unsupported calls
  • Coding work

    Code generation, repair, and software-engineering tasks

    Not enough matched evidence

    Hy4 preview is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.

    Confidence: limited
  • Long documents

    Prompts that approach the documented context limit

    No clear pick

    The documented context windows are equal.

    Confidence: documented
  • Chat turn cost

    1K fresh input + 500 output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: listed-rates
  • Cache-heavy agent loop cost

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

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: listed-rates
  • Repository review cost

    50K fresh input + 3K output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    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.

71.7Hy4 preview63.2Mistral Large 4

Directional only · BenchAlign v6.0

Hy4 preview has the higher coding point estimate. Conditional score ranges do not establish rank confidence.

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

Same basis rules as the category table below

Or describe your own job

The LLM Selector reads what you need a model for and returns a shortlist, ranked in the same order the job pages publish.

Describe the task and constraints. This selector recommends models; it cannot carry out the task or access your files and accounts.

We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy

What is actually comparable

Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.

3 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.

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.

A shared-evidence shape is not available.

BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.

Category results, on a stated basis

Each row shows the public-lane category score for both models: the BenchAlign v6.0 lane in every category, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence. Directional and not-comparable rows remain visible, but they never receive a winner.

Agentic

Like-for-like
Hy4 preview
66.8
Supported · #31/179
Mistral Large 4
56.2
Supported · #74/179
Basis
BenchAlign v6.0 lane · 18 vs 8 public rows
Reading
Hy4 preview leads · intervals overlap

Reasoning

Like-for-like
Hy4 preview
66.0
Supported · #29/161
Mistral Large 4
56.0
Supported · #57/161
Basis
BenchAlign v6.0 lane · 2 vs 1 public rows
Reading
Hy4 preview leads · intervals overlap

Coding

Directional only
Hy4 preview
71.7
Estimated · #24/179
Mistral Large 4
63.2
Supported · #43/179
Basis
BenchAlign v6.0 lane · 8 vs 2 public rows
Reading
Directional only

Multimodal

Directional only
Hy4 preview
67.9
Estimated · #27/119
Mistral Large 4
50.8
Estimated · #77/119
Basis
BenchAlign v6.0 lane · 1 vs 0 public rows
Reading
Directional only

Knowledge

Directional only
Hy4 preview
63.3
Estimated · #33/158
Mistral Large 4
48.1
Estimated · #96/158
Basis
BenchAlign v6.0 lane · 4 vs 0 public rows
Reading
Directional only

Multilingual

Not comparable
Hy4 preview
62.7
Estimated · #15/87
Mistral Large 4
Not ranked
Basis
BenchAlign v6.0 lane · 0 vs 0 public rows
Reading
Not comparable

Instruction following

Not comparable
Hy4 preview
Not ranked
Mistral Large 4
53.5
Estimated · #62/143
Basis
BenchAlign v6.0 lane · 0 vs 0 public rows
Reading
Not comparable

Math

Not comparable
Hy4 preview
Not ranked
Mistral Large 4
55.5
Estimated · #48/125
Basis
BenchAlign v6.0 lane · 1 vs 0 public rows
Reading
Not comparable

Ranks count the models scored in each category’s BenchAlign v6.0 lane, so denominators differ by category. 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 preview
Self-hosted; infrastructure cost varies
Fits in one request
Mistral Large 4
$0.00175
Fits in one request

Hy4 preview has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

Hy4 preview
Self-hosted; infrastructure cost varies
Fits in one request
Mistral Large 4
$0.0413
Fits in one request

Hy4 preview has no comparable published API token rate.

Cache-heavy agent loop

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

Hy4 preview
Self-hosted; infrastructure cost varies
Fits in one request
Cached-input rate unavailable
Mistral Large 4
$0.049
Fits in one request

Hy4 preview has no comparable published API token rate.

Cached input falls back to the list input rate only where a cached rate is unpublished

Price against score, among all priced models

Overall score against output price per 1M tokens, one model per family. Pricing as of November 20, 2026.

Mistral Large 4

20406080100$0.10$0.50$1$5$10$50$100Output price, $ per 1M tokens (log)Overall score
  • Hy4 preview has no public API price, so it is not plotted.
  • 5 cheaper models score at least as high as Mistral Large 4. The nearest is DeepSeek V4.1 Flash 1017 (score 70.0, $1.20 per 1M output tokens).

Gray dots are the other priced models. The dark step line is the efficiency frontier: for a model under it, another model costs no more and scores at least as high.

See Mistral Large 4 on the full price vs performance chart

Specification differences

Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.

Provider availability

Hy4 preview

Not sourced

Mistral Large 4

Not sourced

Reasoning profile

Hy4 preview

Reasoning

Mistral Large 4

Hybrid

Weight access

Hy4 preview

Open Weight

Mistral Large 4

Open Weight

License

Hy4 preview

Open Weight

Mistral Large 4

Open Weight

Release date

Hy4 preview

2026-08-28

Mistral Large 4

2026-10-06

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
Hy4 preview has the higher public point estimate, 68.2 versus 56.9. Their conditional score ranges overlap. These ranges do not establish rank confidence.
Workload cost
A complete comparable API-rate estimate is not available for both models.
Context tradeoff
Both models list 1M.
Run the same representative tasks against both endpoints before changing production traffic.

Questions

Which is better, Hy4 preview or Mistral Large 4?

Hy4 preview has the higher public point estimate, 68.2 versus 56.9. Their conditional score ranges overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.

Which is better for coding, Hy4 preview or Mistral Large 4?

Hy4 preview scores higher for coding on the public lane, 71.7 to 63.2. Hy4 preview 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, Hy4 preview or Mistral Large 4?

Hy4 preview has the higher public agentic tasks point estimate, 66.8 to 56.2, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

Which costs less, Hy4 preview or Mistral Large 4?

Both models do not have comparable published API token rates, so this page does not name a universal price winner.

Which has the larger context window, Hy4 preview or Mistral Large 4?

Both models list the same context window, 1M.

Self-host vs API cost

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

Hy4 preview
API / mo (Novita AI)$2,501
Self-host / mo$27,962
Break-evenNot reached at 70% load
Mistral Large 4
API / mo$2,100
Self-host / mo$37,843
Break-evenNot reached at 70% load
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 evidence43 rows

Agentic

  • Terminal-Bench 2.1

    Hy4 preview85.4%
    Source
    Mistral Large 4—

    Not directly comparable

  • CyberGym

    Hy4 preview78.4%
    Source
    Mistral Large 4—

    Not directly comparable

  • WideResearch

    Hy4 preview83.9%
    Source
    Mistral Large 4—

    Not directly comparable

  • DRACO

    Hy4 preview77.2%
    Source
    Mistral Large 4—

    Not directly comparable

  • MCP Atlas

    Hy4 preview83.7%
    Source
    Mistral Large 4—

    Not directly comparable

  • Toolathlon-Verified

    Hy4 preview74.1%
    Source
    Mistral Large 4—

    Not directly comparable

  • APEX-Agents

    Hy4 preview37.1%
    Source
    Mistral Large 4—

    Not directly comparable

  • skillsBench

    Hy4 preview62.9%
    Source
    Mistral Large 4—

    Not directly comparable

  • JobBench

    Hy4 preview61.7%
    Source
    Mistral Large 4—

    Not directly comparable

  • Agents' Last Exam

    Hy4 preview22.8%
    Source
    Mistral Large 4—

    Not directly comparable

  • AutomationBench

    Hy4 preview32.1%
    Source
    Mistral Large 4—

    Not directly comparable

  • BankerToolBench

    Hy4 preview78.6%
    Source
    Mistral Large 4—

    Not directly comparable

  • HLE w/ tools

    Hy4 preview55.4%
    Source
    Mistral Large 4—

    Not directly comparable

  • Terminal-Bench 2.1 (Vals)

    Hy4 preview55.1%
    Source
    Mistral Large 4—

    Not directly comparable

  • CWE-bench v1

    Hy4 preview53.0%
    Source
    Mistral Large 4—

    Not directly comparable

  • Hermes Index

    Hy4 preview41.0%
    Source
    Mistral Large 4—

    Not directly comparable

  • Hermes Bench

    Hy4 preview70.8%
    Source
    Mistral Large 4—

    Not directly comparable

  • Hy4 preview47.3%
    Mistral Large 445.6%

    Hy4 preview leads this result

  • Terminal-Bench 4.1

    Hy4 preview—
    Mistral Large 427.60%
    Source

    Not directly comparable

  • Finch

    Hy4 preview—
    Mistral Large 473.3%
    Source

    Not directly comparable

  • b3 Agent Security

    Hy4 preview—
    Mistral Large 497.1%
    Source

    Not directly comparable

  • Cyber Misuse Refusal

    Hy4 preview—
    Mistral Large 496.6%
    Source

    Not directly comparable

  • Cybench

    Hy4 preview—
    Mistral Large 493.0%
    Source

    Not directly comparable

  • BrowseComp

    Hy4 preview—
    Mistral Large 484.5%
    Source

    Not directly comparable

  • Finance Agent v2

    Hy4 preview—
    Mistral Large 454.7%
    Source

    Not directly comparable

Coding

  • Terminal-Bench 2.1

    Hy4 preview85.4%
    Source
    Mistral Large 4—

    Not directly comparable

  • SWE-bench Pro

    Hy4 preview65.7%
    Source
    Mistral Large 4—

    Not directly comparable

  • SWE Multilingual

    Hy4 preview82.9%
    Source
    Mistral Large 4—

    Not directly comparable

  • DeepSWE

    Hy4 preview64.3%
    Source
    Mistral Large 4—

    Not directly comparable

  • NL2Repo

    Hy4 preview58.9%
    Source
    Mistral Large 4—

    Not directly comparable

  • ProgramBench

    Hy4 preview17.5%
    Source
    Mistral Large 4—

    Not directly comparable

  • PostTrainBench (lab runs)

    Hy4 preview35.6%
    Source
    Mistral Large 4—

    Not directly comparable

  • sweMarathon

    Hy4 preview31.9%
    Source
    Mistral Large 4—

    Not directly comparable

  • Vibe Code Bench

    Hy4 preview—
    Mistral Large 478.40%
    Source

    Not directly comparable

  • FrontierCode 2.0

    Hy4 preview—
    Mistral Large 419.7%
    Source

    Not directly comparable

Reasoning

  • Hy4 preview2.59
    Mistral Large 42.50

    Hy4 preview leads this result

  • CritPt

    Hy4 preview16.9%
    Source
    Mistral Large 4—

    Not directly comparable

Multimodal

  • OfficeQA Pro

    Hy4 preview66.2%
    Source
    Mistral Large 4—

    Not directly comparable

Knowledge

  • GPQA

    Hy4 preview92.3%
    Source
    Mistral Large 4—

    Not directly comparable

  • GPQA-D

    Hy4 preview92.3%
    Source
    Mistral Large 4—

    Not directly comparable

  • HLE

    Hy4 preview55.4%
    Source
    Mistral Large 4—

    Not directly comparable

  • HLE w/o tools

    Hy4 preview43.4%
    Source
    Mistral Large 4—

    Not directly comparable

Math

  • Apex

    Hy4 preview74.2%
    Source
    Mistral Large 4—

    Not directly comparable

43 public results · 2 shared

Watch Hy4 preview vs Mistral Large 4

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

Read a sample issue

Join 7,000+ subscribers.

Last updated November 21, 2026