Noometry

Model comparison

GPT-5 Nano vs Hy3

Hy3 is the stronger model overall, scoring 44.2 to 33.5 on the Noometry Index.

Last verified . 16 shared benchmarks.

GPT-5 Nano OpenAI

33.5

Rank #241 Confirmed

Hy3 Tencent

44.2

Rank #79 Confirmed

Summary

  • They share 16 benchmarks with published results for both. GPT-5 Nano scores higher in 0 categories and Hy3 in 8 categories; 7 gaps are clear of the uncertainty.
  • The widest gap is in writing & preference, where Hy3 leads 62.2 to 39.1.
  • Both cost about the same: $0.05 input and $0.40 output per million tokens.
  • GPT-5 Nano accepts more context: 400K tokens versus 262K.
  • Hy3 has downloadable open weights; the other is API-only.

Side by side

GPT-5 NanoHy3
ProviderOpenAITencent
Noometry Index33.544.2
Released2025-08-072026-07-06
WeightsProprietaryOpen
Context window400K262K
Max output128K128K
Input $ / M tokens$0.05$0.0825
Output $ / M tokens$0.40$0.33
Results tracked4919

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Hy3 leads

GPT-5 Nano: 33.6 (#254), Hy3: 46.8 (#63)

BenchmarkGPT-5 NanoHy3
LMArena Coding13511464
SWE-bench Verified (bash only)34.8%—
LMArena WebDev—1508
WeirdML38.1%—
ALE-Bench718.67—

Agentic & Tool Use Not comparable

GPT-5 Nano: 25.8 (#106), Hy3: —

BenchmarkGPT-5 NanoHy3
Terminal-Bench21.8%—
Berkeley Function Calling Leaderboard51.5%—

Reasoning Hy3 leads

GPT-5 Nano: 16.3 (#306), Hy3: 26.1 (#136)

BenchmarkGPT-5 NanoHy3
LMArena Hard Prompts13281447
ARC-AGI-22.6%—
Kagi LLM Benchmark62.2%—
NYT Connections (extended)—41.2%
ARC-AGI-120.7%—
Chess Puzzles27%—
Mystery Game Puzzles9%—
DTBench62.7%—
LMCA7.9%—
Epoch Capabilities Index139.38—
ForecastBench59.1—

Math Hy3 leads

GPT-5 Nano: 29.4 (#241), Hy3: 40.1 (#93)

BenchmarkGPT-5 NanoHy3
LMArena Math13171475
FrontierMath (Tiers 1-3)20%—
FrontierMath Tier 42.4%—
OTIS Mock AIME 2024-202581.1%—
ProofBench12%—
Omni-MATH54.6%—
MATH Level 595.2%—
FrontierMath (Feb 2025 set)8.3%—
FrontierMath Tier 4 (v1)2.1%—

Knowledge Hy3 leads

GPT-5 Nano: 35.9 (#178), Hy3: 40.8 (#114)

BenchmarkGPT-5 NanoHy3
LMArena Expert13211460
GPQA Diamond69.4%—
SimpleQA Verified11.7%—
MMLU-Pro77.8%—
Vectara Hallucination Rate10.5%—
GPQA (HELM)67.9%—

Multimodal Not comparable

GPT-5 Nano: 31.3 (#108), Hy3: —

BenchmarkGPT-5 NanoHy3
LMArena Vision1159—
VPCT37.2%—

Multilingual Hy3 leads

GPT-5 Nano: 45.3 (#172), Hy3: 53.5 (#65)

BenchmarkGPT-5 NanoHy3
LMArena Non-English13131426
LMArena Chinese13561493
LMArena German13271439
LMArena Japanese12261392
LMArena Korean12691395
LMArena Russian12961432
LMArena Spanish13601456
LMArena French—1461

Instruction Following Too close to call

GPT-5 Nano: 75.0 (#79), Hy3: 75.1 (#70)

BenchmarkGPT-5 NanoHy3
LMArena Instruction Following13061426
IFEval93.2%—

Long Context Hy3 leads

GPT-5 Nano: 31.3 (#281), Hy3: 44.1 (#75)

BenchmarkGPT-5 NanoHy3
LMArena Longer Query13121442
Fiction.LiveBench44.4%—

Writing & Preference Hy3 leads

GPT-5 Nano: 39.1 (#249), Hy3: 62.2 (#81)

BenchmarkGPT-5 NanoHy3
LMArena Text13201439
LMArena Creative Writing12491402
LMArena Multi-Turn13111436
EQ-Bench Creative Writing705—
WildBench80.6%—

Frequently asked questions

Is GPT-5 Nano better than Hy3?

Hy3 is the stronger model overall, scoring 44.2 to 33.5 on the Noometry Index.

Which is cheaper, GPT-5 Nano or Hy3?

GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Hy3 lists at $0.0825 and $0.33.

Is GPT-5 Nano or Hy3 better for coding?

Hy3 scores higher on coding benchmarks: 46.8 versus 33.6 in the Noometry coding category.

Which has the bigger context window?

GPT-5 Nano does, with 400K tokens against 262K.

How many benchmarks do GPT-5 Nano and Hy3 share?

16 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and Hy3 has 19.

Related comparisons

Go deeper