VerdictPal
Join

Models / OpenBMB

MiniCPM5-2B

OpenBMB · Released 2026-09-07

AA Intelligence Index 14.3. Listed on Artificial Analysis with API pricing $0/$0 per 1M tokens. Benchmark rows sync from the AA snapshot — see the ledger on this page.

Small OpenBMB checkpoint tracked on Artificial Analysis. Use the ledger below for sourced scores; we do not infer numbers beyond published rows.

#161 of 184 on AA Index · current snapshot
14 AA Index · 9 benchmark rows · 1 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite29
ContextUnlisted
Input / 1MFree
Output / 1MFree
Knowledge cutoffunpublished
Statuslightweight

Family profile

Best published score in each covered benchmark family.

44/100 avg
Reasoning
70
Coding
26
Agentic
21
Long context
59

4 tested benchmark families

Benchmark placements

Where MiniCPM5-2B places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
τ³-Bench BankingAgentic#63/ 11121Artificial Analysis2026-09-10
SciCodeCoding#91/ 9426Artificial Analysis2026-09-10
AA-LCRLong context#149/ 18259Artificial Analysis2026-09-10
Artificial Analysis Intelligence IndexReasoning#161/ 18814Artificial Analysis2026-09-10
Humanity's Last ExamReasoning#162/ 1859Artificial Analysis2026-09-10
GPQA DiamondReasoning#170/ 19070Artificial Analysis2026-09-10
AA output speedPerformance#174/ 190Artificial Analysis2026-09-10
AA time to first tokenPerformance#174/ 190Artificial Analysis2026-09-10
Terminal-BenchAgentic#181/ 1899Artificial Analysis2026-09-10
Family context

The three highest-scoring models with pages in each capability family. Where MiniCPM5-2B shows up, it's highlighted.

ReasoningMiniCPM5-2B · 70
  1. 1
    GPT-6 Astra
    96
  2. 2
    Claude Mythos Preview
    95
  3. 3
    Gemini 3.7 Flash
    95
CodingMiniCPM5-2B · 26
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticMiniCPM5-2B · 21
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextMiniCPM5-2B · 59
  1. 1
    Kimi K3
    89
  2. 2
    Claude Fable 5.1
    85
  3. 3
    GPT-5.5 (high)
    84
Newest receipts

Benchmark rows added to the public ledger for MiniCPM5-2B in the last 120 days. Older rows live in the full table below.

9 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context59.33%Artificial Analysis4
AA output speedPerformance0 t/sArtificial Analysis4
AA time to first tokenPerformance0.00sArtificial Analysis4
Artificial Analysis Intelligence IndexReasoning14.3%Artificial Analysis4
GPQA DiamondReasoning70.2%Artificial Analysis4
Humanity's Last ExamReasoning8.9%Artificial Analysis4
SciCodeCoding26.3%Artificial Analysis4
τ³-Bench BankingAgentic20.82%Artificial Analysis4
Terminal-BenchAgentic8.61%Artificial Analysis4
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Reasoning70
Coding26
Agentic21
Long context59
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMetaxAIMiniMaxZhipuXiaomiOtherMeituanCohereNVIDIA

Full benchmark ledger

Every catalog benchmark for MiniCPM5-2B. Scores link to the original source; gaps mean no public row exists yet.

9 sourced rows

9 of 37 catalog benchmarks have a sourced row for MiniCPM5-2B.

Reasoning

Coding

Agentic

Long context

Performance

Sources