Skip to main content
BenchLM
Data

Gemini 3 Pro vs Step 5 Preview

Updated November 5, 2026. Rank says Gemini 3 Pro 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

Gemini 3 Pro has the higher public point estimate, 60.94 versus 56.81. 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
Google logo
Gemini 3 Pro

Google

60.94/100

Supported · Public rank #51

90% interval 49.8–72.1

Model B
StepFun logo
Step 5 Preview

StepFun

56.81/100

Estimated · Public rank #60

Conditional range 47.1–66.5

Shared results
2
Gemini 3 Pro only
25
Step 5 Preview only
22
Like-for-like categories
0 / 8
Supported: Gemini 3 Pro · Estimated: Step 5 Preview. 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.

  • Long documents

    Prompts that approach the documented context limit

    Gemini 3 Pro

    Gemini 3 Pro has the larger documented context window.

    Confidence: documented

  • Chat turn cost

    1K fresh input + 500 output tokens

    Step 5 Preview

    Step 5 Preview 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

    Step 5 Preview

    Step 5 Preview has the lower estimated token cost for this stated workload. Gemini 3 Pro has no published cached-input rate, so cached tokens use its listed input rate.

    Confidence: rate-fallback
Show secondary and unsupported calls

  • Repository review cost

    50K fresh input + 3K output tokens

    Step 5 Preview

    Step 5 Preview 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

    Gemini 3 Pro and Step 5 Preview are 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

    Gemini 3 Pro is not ranked on the public lane for agentic, so no winner is named for agentic.

    Confidence: limited

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.

43.9Gemini 3 Pro48.8Step 5 Preview

Directional only · BenchAlign v5.8

Step 5 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.

Open the coding leaderboardShortlist models for this job
Same basis rules as the category table below

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.

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.

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.

Coding

Directional only
Gemini 3 Pro
43.9
Estimated · #60/156
Step 5 Preview
48.8
Estimated · #47/156
Basis
BenchAlign v5.8 lane · 1 vs 6 public rows
Reading
Directional only

Multimodal

Directional only
Gemini 3 Pro
79.6
#22/58
Step 5 Preview
73.3
#32/58
Basis
Provisional lane · 2 vs 2 weighted rows
Reading
Directional only

Knowledge

Directional only
Gemini 3 Pro
60.6
Estimated · #45/177
Step 5 Preview
61.0
Supported · #44/177
Basis
BenchAlign v5.8 lane · 0 vs 2 public rows
Reading
Directional only

Agentic

Not comparable
Gemini 3 Pro
Not ranked
Step 5 Preview
55.3
Estimated · #42/135
Basis
BenchAlign v5.8 lane · 2 vs 13 public rows
Reading
Not comparable

Reasoning

Not comparable
Gemini 3 Pro
50.1
Unranked · 4 rankable rows
Step 5 Preview
82.5
Unranked · 2 rankable rows
Basis
Provisional lane · 1 vs 0 weighted rows
Reading
Not comparable

Multilingual

Not comparable
Gemini 3 Pro
Not ranked
Step 5 Preview
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
Not comparable

Instruction following

Not comparable
Gemini 3 Pro
86.1
#48/157
Step 5 Preview
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
Not comparable

Math

Not comparable
Gemini 3 Pro
Not ranked
Step 5 Preview
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
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

Gemini 3 Pro
$0.008
Fits in one request
Step 5 Preview
$0.00235
Fits in one request

Step 5 Preview has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

Gemini 3 Pro
$0.136
Fits in one request
Step 5 Preview
$0.0581
Fits in one request

Step 5 Preview has the lower modeled cost

Costs use the listed standard API rates.

Cache-heavy agent loop

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

Gemini 3 Pro
$0.56
Fits in one request
Cached input priced at the published list-input rate
Step 5 Preview
$0.057
Fits in one request

Step 5 Preview has the lower modeled cost

Gemini 3 Pro has no published cached-input rate, so cached tokens use its listed input rate.

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.

Cached-input rate

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

Gemini 3 Pro

Not published

Step 5 Preview

$0.05 per 1M cached input tokens

StepFun pricing and rate limits

Reasoning profile

Gemini 3 Pro

Non-Reasoning

Step 5 Preview

Reasoning

Weight access

Gemini 3 Pro

Proprietary

Step 5 Preview

Open Weight

License

Gemini 3 Pro

Proprietary

Step 5 Preview

Open Weight

Release date

Gemini 3 Pro

2025-11-18

Step 5 Preview

2026-09-20

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
Gemini 3 Pro has the higher public point estimate, 60.94 versus 56.81. Their conditional score ranges overlap. These ranges do not establish rank confidence.
Workload cost
Repository review: $0.136 vs $0.0581. Cache-heavy agent loop: $0.56 vs $0.057.
Context tradeoff
Gemini 3 Pro has the larger documented window (2M).
Run the same representative tasks against both endpoints before changing production traffic.

Questions

Which is better, Gemini 3 Pro or Step 5 Preview?

Gemini 3 Pro has the higher public point estimate, 60.94 versus 56.81. 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, Gemini 3 Pro or Step 5 Preview?

Step 5 Preview scores higher for coding on the public lane, 48.8 to 43.9. Gemini 3 Pro and Step 5 Preview are 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, Gemini 3 Pro or Step 5 Preview?

Gemini 3 Pro is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.

Which costs less, Gemini 3 Pro or Step 5 Preview?

For the stated presets, chat costs $0.008 on Gemini 3 Pro and $0.00235 on Step 5 Preview; repository review costs $0.136 and $0.0581; the cache-heavy agent loop costs $0.56 and $0.057. Gemini 3 Pro has no published cached-input rate, so cached tokens use its listed input rate.

Which has the larger context window, Gemini 3 Pro or Step 5 Preview?

Gemini 3 Pro has the larger documented context window: 2M, compared with 1M.

Self-host vs API cost

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

Gemini 3 Pro
API / mo$10,500
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
Step 5 Preview
API / mo$2,775
Self-host / mo$27,962
Break-even617M/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 evidence49 rows

Agentic

  • Gert Labs

    Gemini 3 Pro63.23%
    Source
    Step 5 Preview—

    Not directly comparable

  • JobBench

    Gemini 3 Pro11.4%
    Source
    Step 5 Preview59.0%
    Source

    Step 5 Preview leads this result

  • Terminal-Bench 2.1

    Gemini 3 Pro—
    Step 5 Preview85.0%
    Source

    Not directly comparable

  • Terminal-Bench 4.0

    Gemini 3 Pro—
    Step 5 Preview33.30%
    Source

    Not directly comparable

  • CyberGym

    Gemini 3 Pro—
    Step 5 Preview84.7%
    Source

    Not directly comparable

  • AutomationBench

    Gemini 3 Pro—
    Step 5 Preview44.0%
    Source

    Not directly comparable

  • Toolathlon-Verified

    Gemini 3 Pro—
    Step 5 Preview74.1%
    Source

    Not directly comparable

  • MCP Atlas

    Gemini 3 Pro—
    Step 5 Preview85.6%
    Source

    Not directly comparable

  • APEX-Agents

    Gemini 3 Pro—
    Step 5 Preview37.8%
    Source

    Not directly comparable

  • DRACO

    Gemini 3 Pro—
    Step 5 Preview83.3%
    Source

    Not directly comparable

  • BrowseComp

    Gemini 3 Pro—
    Step 5 Preview88.7%
    Source

    Not directly comparable

  • Agents' Last Exam

    Gemini 3 Pro—
    Step 5 Preview29.5%
    Source

    Not directly comparable

  • Hermes Index

    Gemini 3 Pro—
    Step 5 Preview32.8%
    Source

    Not directly comparable

  • Hermes Bench

    Gemini 3 Pro—
    Step 5 Preview69.2%
    Source

    Not directly comparable

Coding

  • Vibe Code Bench

    Gemini 3 Pro14.30%
    Source
    Step 5 Preview—

    Not directly comparable

  • DeepSWE

    Gemini 3 Pro—
    Step 5 Preview67.7%
    Source

    Not directly comparable

  • Terminal-Bench 2.1

    Gemini 3 Pro—
    Step 5 Preview85.0%
    Source

    Not directly comparable

  • SciCode

    Gemini 3 Pro—
    Step 5 Preview58.9%
    Source

    Not directly comparable

  • ProgramBench

    Gemini 3 Pro—
    Step 5 Preview80.5%
    Source

    Not directly comparable

  • sweMarathon

    Gemini 3 Pro—
    Step 5 Preview72.7%
    Source

    Not directly comparable

  • MLS-Bench Lite

    Gemini 3 Pro—
    Step 5 Preview40.5%
    Source

    Not directly comparable

Reasoning

  • ARC-AGI-2

    Gemini 3 Pro31.1%
    Source
    Step 5 Preview—

    Not directly comparable

  • ARC-AGI-1

    Gemini 3 Pro75.00%
    Source
    Step 5 Preview—

    Not directly comparable

  • CritPt

    Gemini 3 Pro—
    Step 5 Preview20.9%
    Source

    Not directly comparable

Multimodal

  • MMMU

    Gemini 3 Pro87.2%
    Source
    Step 5 Preview—

    Not directly comparable

  • MMMU-Pro

    Gemini 3 Pro81%
    Source
    Step 5 Preview76%
    Source

    Gemini 3 Pro leads this result

  • RealWorldQA

    Gemini 3 Pro83.3%
    Source
    Step 5 Preview—

    Not directly comparable

  • OmniDocBench 1.5

    Gemini 3 Pro88.5%
    Source
    Step 5 Preview—

    Not directly comparable

  • Video-MME (with subtitle)

    Gemini 3 Pro88.4%
    Source
    Step 5 Preview—

    Not directly comparable

  • Video-MME (w/o subtitle)

    Gemini 3 Pro87.7%
    Source
    Step 5 Preview—

    Not directly comparable

  • MathVision

    Gemini 3 Pro86.6%
    Source
    Step 5 Preview—

    Not directly comparable

  • We-Math

    Gemini 3 Pro86.9%
    Source
    Step 5 Preview—

    Not directly comparable

  • DynaMath

    Gemini 3 Pro85.1%
    Source
    Step 5 Preview—

    Not directly comparable

  • SimpleVQA

    Gemini 3 Pro73.2%
    Source
    Step 5 Preview—

    Not directly comparable

  • MMLongBench-Doc

    Gemini 3 Pro60.5%
    Source
    Step 5 Preview—

    Not directly comparable

  • CC-OCR

    Gemini 3 Pro79.0%
    Source
    Step 5 Preview—

    Not directly comparable

  • AI2D_TEST

    Gemini 3 Pro94.1%
    Source
    Step 5 Preview—

    Not directly comparable

  • CountBench

    Gemini 3 Pro97.3%
    Source
    Step 5 Preview—

    Not directly comparable

  • RefCOCO (avg)

    Gemini 3 Pro84.1%
    Source
    Step 5 Preview—

    Not directly comparable

  • ODINW13

    Gemini 3 Pro46.3%
    Source
    Step 5 Preview—

    Not directly comparable

  • ERQA

    Gemini 3 Pro70.5%
    Source
    Step 5 Preview—

    Not directly comparable

  • VideoMMMU

    Gemini 3 Pro87.6%
    Source
    Step 5 Preview—

    Not directly comparable

  • MLVU (M-Avg)

    Gemini 3 Pro83.0%
    Source
    Step 5 Preview—

    Not directly comparable

  • ScreenSpot Pro

    Gemini 3 Pro72.7%
    Source
    Step 5 Preview—

    Not directly comparable

  • CharXiv

    Gemini 3 Pro81.4%
    Source
    Step 5 Preview—

    Not directly comparable

  • V*

    Gemini 3 Pro88.0%
    Source
    Step 5 Preview—

    Not directly comparable

  • OfficeQA Pro

    Gemini 3 Pro—
    Step 5 Preview60.3%
    Source

    Not directly comparable

Knowledge

  • GPQA-D

    Gemini 3 Pro—
    Step 5 Preview93.5%
    Source

    Not directly comparable

  • HLE

    Gemini 3 Pro—
    Step 5 Preview46.5%
    Source

    Not directly comparable

49 public results · 2 shared

Related comparisons

Compare API pricingRead the methodologyOpen the model selector

Watch Gemini 3 Pro vs Step 5 Preview

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 5, 2026