llmboard.ai
Benchmarks
CompareRankings
llmboard.ai
Benchmarks
CompareRankings
HomeModelsPhi 3.5 MoE

Microsoft model product

Phi 3.5 MoE

Phi-3.5-MoE-instruct is a mixture-of-experts model with ~42B total parameters (6.6B active) and a 128K context window. It excels at reasoning, math, coding, and multilingual tasks, outperforming larger dense models in many benchmarks. It underwent a thorough safety post-training process (SFT + DPO) and is licensed under MIT. This model is ideal for scenarios where efficiency and high performance are both required, particularly in multi-lingual or reasoning-intensive tasks.

Updated Aug 10, 2026. Default version: Phi-3.5-MoE-instruct

Compare
LLMBoard score42.4Phi-3.5-MoE-instruct
Coverage80%14 benchmark families
Context windowN/ATokens
Official input priceN/AOfficial price unavailable

On this page

  • Specification
  • Capability
  • Benchmarks
  • Arena
  • Pricing
  • Versions
  • About
  • Compare
  • Similar models
  • FAQ

Model specification

Structured fields from the published default version.

Version
Phi-3.5-MoE-instruct
Released
Aug 23, 2024
Knowledge cutoff
Unknown
Parameters
60B
Context window
N/A
Max output
N/A
Inputs
text
Outputs
text
Open weights
No
License
MIT

Capability profile

This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.

Phi-3.5-MoE-instruct category scores

Benchmark results

Published benchmark records for the scored version Phi-3.5-MoE-instruct.

31 rows
Columns

Show columns

GovReport0.312100.0%CAug 7, 2026
MEGA MLQA0.712100.0%CAug 7, 2026
MEGA TyDi QA0.712100.0%CAug 7, 2026
MEGA UDPOS0.612100.0%CAug 7, 2026
MEGA XCOPA0.812100.0%CAug 7, 2026
MEGA XStoryCloze0.812100.0%CAug 7, 2026
OpenBookQA0.915100.0%CAug 7, 2026
PIQA0.9111100.0%CAug 7, 2026
RepoQA0.812100.0%CAug 7, 2026
Social IQa0.819100.0%CAug 7, 2026
SummScreenFD0.212100.0%CAug 7, 2026
Qasper0.4220.0%CAug 7, 2026
QMSum0.2220.0%CAug 7, 2026
SQuALITY0.22575.0%CAug 7, 2026
TruthfulQA0.821894.1%CAug 7, 2026
BoolQ0.831077.8%CAug 7, 2026
RULER0.93433.3%CAug 7, 2026
BIG-Bench Hard0.892160.0%CAug 7, 2026
Winogrande0.892261.9%CAug 7, 2026
ARC-C0.9103472.7%CAug 7, 2026
MBPP0.8123769.4%CAug 7, 2026
HellaSwag0.8142750.0%CAug 7, 2026
Arena Hard0.4222616.0%CAug 7, 2026
MGSM0.6263116.7%CAug 7, 2026
GSM8k0.9304838.3%CAug 7, 2026
MMMLU0.7444910.4%CAug 7, 2026
MATH0.6497131.4%CAug 7, 2026
HumanEval0.7647616.0%CAug 7, 2026
MMLU0.86910031.3%CAug 7, 2026
MMLU-Pro0.51201297.0%CAug 7, 2026
GPQA0.42102339.9%CAug 7, 2026

Arena results

Preference and agent-evaluation signals from published Arena datasets.

No published Arena match

The default version has no published Arena rows, or its source alias has not been resolved.

Pricing

Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.

Official API
N/A
Official provider
N/A
Lowest third-party
N/A
Tracked offerings
0
No published price snapshot

The default version has no provider offering with current input or output token prices.

Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.

Phi 3.5 MoE versions

All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.

1 rows
Columns

Show columns

Phi-3.5-MoE-instructAug 23, 202442.460BN/AN/ANoMIT

What is Phi 3.5 MoE?

A concise description based on the published model registry.

Phi-3.5-MoE-instruct is a mixture-of-experts model with ~42B total parameters (6.6B active) and a 128K context window. It excels at reasoning, math, coding, and multilingual tasks, outperforming larger dense models in many benchmarks. It underwent a thorough safety post-training process (SFT + DPO) and is licensed under MIT. This model is ideal for scenarios where efficiency and high performance are both required, particularly in multi-lingual or reasoning-intensive tasks.

Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.

Data snapshot: 2026-08-07. Editorial model content is not available in the backend.

Phi 3.5 MoE vs nearby models

Open a comparison with the three ranked models immediately above and below this model.

Phi 3.5 MoEvsMistral Small 3.1 24BPhi 3.5 MoEvsGPT-5.4-nanoPhi 3.5 MoEvsQwen3 VL 235B A22BPhi 3.5 MoEvsQwen3 Next 80B A3BPhi 3.5 MoEvsGPT-4o-miniPhi 3.5 MoEvsClaude Haiku 3.5

Models similar to Phi 3.5 MoE

Recommendations prioritize the same model type and family, then the closest published LLMBoard score.

#121-5.5
MI

Phi 4

Microsoft

36.9 LLMBoard

DetailsCompare
#92+5.5
MI

Phi 4 Reasoning

Microsoft

47.9 LLMBoard

DetailsCompare
#83+8.0
MI

Phi 4 Reasoning Plus

Microsoft

50.3 LLMBoard

DetailsCompare
#134-10.9
MI

Phi 4 Mini

Microsoft

31.5 LLMBoard

DetailsCompare
#144-16.5
MI

Phi 3.5 mini

Microsoft

25.9 LLMBoard

DetailsCompare
#44+21.0
MI

MAI Thinking 1

Microsoft

63.4 LLMBoard

DetailsCompare

FAQ

Common questions about Phi 3.5 MoE.

When was Phi 3.5 MoE released?

Phi 3.5 MoE's default version was released on Aug 23, 2024.

How much does Phi 3.5 MoE cost?

No official standard PAYG price is currently available for Phi 3.5 MoE.

Who created Phi 3.5 MoE?

Phi 3.5 MoE is published under Microsoft in the model registry.

What is the context window for Phi 3.5 MoE?

The current registry does not publish a context window for the default version.

Is Phi 3.5 MoE open weight?

No. The default version is not marked as having publicly available weights.

How many API providers offer Phi 3.5 MoE?

No published provider offering is currently linked to the default version.

What models should I compare Phi 3.5 MoE with?

Nearby ranked alternatives include Mistral Small 3.1 24B, GPT-5.4-nano, Qwen3 VL 235B A22B.

Rankings

OverallCodingText ArenaPricing

Modalities

Image GenerationVideo GenerationSpeech-to-TextEmbeddings

Benchmarks

All BenchmarksReasoningMathCoding

Vendors

All VendorsOpenAIAnthropicGoogle
llmboard.aiCopyright 2026 llmboard.ai