ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

MiMo-V2.5vsPhi-4

MiMo-V2.5 leads 1 of 1 shared benchmark. MiMo-V2.5 has the larger context window (1M tokens).

Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.

Shared benchmarks
1 – 0
MiMo-V2.5 leads
Cheaper per token
list price, input + output
Larger context
MiMo-V2.5
1M tokens
Provider uptime (30d)
— · —
Xiaomi MiMo · Microsoft
MiMo-V2.5
Xiaomi MiMo
textvision
Context
1M
Max out
Input /1M
Output /1M
Cached /1M
Scores
7 · 1 core
Phi-4
Microsoft · released 2024-12-12
text
Context
16K
Max out
16K
Input /1M
Output /1M
Cached /1M
Scores
11 · 7 core
Quality

Benchmark matrix

BenchmarkMiMo-V2.5Phi-4ΔEdge
Reported by both · 1
LMArena Elo1427.4 1216.6 +210.8MiMo-V2.5
Only MiMo-V2.5 reports · 6
ChartQA81.4% not reported
DocVQA93.2% not reported
InfoVQA72.7% not reported
MMMU55.1% not reported
MMMU-Pro38.5% not reported
ScienceQA Visual97.5% not reported
Only Phi-4 reports · 9
GPQA Diamondnot reported56.1%
MMLUnot reported84.8%
MMLU-Pronot reported70.4%
HumanEvalnot reported82.6%
ArenaHardnot reported75.4%
HumanEval+not reported82.8%
IFEvalnot reported63%
MATHnot reported80.4%
OTIS Mock AIME 2024-2025 (Epoch AI run)not reported13.8% epoch
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is MiMo-V2.5 minus Phi-4 in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing

Side by side

SpecMiMo-V2.5Phi-4Edge
Context window1M tokens16K tokensMiMo-V2.5
Max output16K tokens
Input price / 1M
Output price / 1M
Cached input / 1M
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
Modalitiestext · visiontextMiMo-V2.5
Released2024-12-12
Cited benchmark scores711
Reliability

Provider status

All providers
Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run MiMo-V2.5 and Phi-4 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.