Model comparison
GPT-5.4 nano vs Kimi K2.5
Kimi K2.5 is the stronger model overall, scoring 47.8 to 41.9 on the Noometry Index. GPT-5.4 nano costs 1.9× less per token, which makes it the better buy when Kimi K2.5's lead doesn't matter for your workload.
Last verified . 33 shared benchmarks.
Summary
- They share 33 benchmarks with published results for both. GPT-5.4 nano scores higher in 0 categories and Kimi K2.5 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Kimi K2.5 leads 53.6 to 41.8.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 39.7% for GPT-5.4 nano and 78.5% for Kimi K2.5.
- GPT-5.4 nano is cheaper at $0.20 / $1.25 per million input/output tokens, against $0.45 / $2.25 for Kimi K2.5.
- GPT-5.4 nano accepts more context: 400K tokens versus 262K.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 nano | Kimi K2.5 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 41.9 | 47.8 |
| Released | 2026-03-17 | 2026-01-27 |
| Weights | Proprietary | Open |
| Context window | 400K | 262K |
| Max output | 128K | 262K |
| Input $ / M tokens | $0.20 | $0.45 |
| Output $ / M tokens | $1.25 | $2.25 |
| Results tracked | 40 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.5 leads
GPT-5.4 nano: 43.1 (#82), Kimi K2.5: 48.7 (#48)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| SciCode | 46.9% | 49% |
| WeirdML | 49.2% | 45.6% |
| LMArena Coding | 1404 | 1474 |
| ALE-Bench | 1,005 | 821.65 |
| SWE-bench Verified | — | 73.8% |
| SWE-bench Verified (bash only) | — | 70.8% |
| LMArena WebDev | — | 1436 |
| SWE-bench Multilingual | — | 67.3% |
Agentic & Tool Use Not comparable
GPT-5.4 nano: —, Kimi K2.5: 34.0 (#44)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | — | 43.2% |
| OSWorld | — | 63.3% |
| Vending-Bench 2 | — | 1,198 |
Reasoning Kimi K2.5 leads
GPT-5.4 nano: 23.8 (#159), Kimi K2.5: 31.2 (#74)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 5.7% | 11.8% |
| Kagi LLM Benchmark | 39.7% | 78.5% |
| ARC-AGI-1 | 51.5% | 65.3% |
| CritPt | 9.3% | 3.1% |
| Chess Puzzles | 30% | 12% |
| LMArena Hard Prompts | 1380 | 1454 |
| Epoch Capabilities Index | 145.78 | 148.05 |
| SimpleBench | — | 46.8% |
| NYT Connections (extended) | — | 69.9% |
| EnigmaEval | — | 3.4% |
| Thematic Generalization | — | 69.4% |
| Mystery Game Puzzles | 9% | — |
| DTBench | 80.3% | — |
| LMCA | 36.9% | — |
| ForecastBench | 57.3 | — |
Math Kimi K2.5 leads
GPT-5.4 nano: 40.7 (#76), Kimi K2.5: 50.3 (#46)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 92.2% |
| LMArena Math | 1406 | 1470 |
| FrontierMath (Feb 2025 set) | 25.9% | 27.9% |
| FrontierMath Tier 4 (v1) | 6.3% | 4.2% |
| FrontierMath (Tiers 1-3) | 44.9% | — |
| FrontierMath Tier 4 | 12.2% | — |
| MathArena Final-Answer Competitions | — | 62.3% |
| ProofBench | 5% | — |
Knowledge Kimi K2.5 leads
GPT-5.4 nano: 41.8 (#94), Kimi K2.5: 53.6 (#52)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 78.5% | 87.6% |
| SimpleQA Verified | 11.7% | 34.3% |
| Vectara Hallucination Rate | 3.1% | 14.2% |
| LMArena Expert | 1397 | 1466 |
| Humanity's Last Exam | — | 24.4% |
Multimodal Kimi K2.5 leads
GPT-5.4 nano: 36.0 (#73), Kimi K2.5: 40.4 (#37)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1196 | 1269 |
| LMArena Document | — | 1430 |
Multilingual Kimi K2.5 leads
GPT-5.4 nano: 48.8 (#136), Kimi K2.5: 54.1 (#50)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1359 | 1433 |
| LMArena Chinese | 1391 | 1495 |
| LMArena French | 1396 | 1454 |
| LMArena German | 1368 | 1441 |
| LMArena Japanese | 1342 | 1422 |
| LMArena Korean | 1321 | 1412 |
| LMArena Russian | 1363 | 1436 |
| LMArena Spanish | 1368 | 1450 |
Instruction Following Kimi K2.5 leads
GPT-5.4 nano: 72.1 (#139), Kimi K2.5: 75.5 (#60)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1362 | 1431 |
Long Context Kimi K2.5 leads
GPT-5.4 nano: 41.6 (#131), Kimi K2.5: 52.1 (#7)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| LMArena Longer Query | 1365 | 1445 |
| Fiction.LiveBench | — | 86.1% |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Kimi K2.5 leads
GPT-5.4 nano: 55.9 (#137), Kimi K2.5: 65.3 (#51)
| Benchmark | GPT-5.4 nano | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1372 | 1445 |
| LMArena Creative Writing | 1313 | 1423 |
| LMArena Multi-Turn | 1382 | 1444 |
| EQ-Bench Creative Writing | — | 1579 |
Frequently asked questions
Is GPT-5.4 nano better than Kimi K2.5?
Kimi K2.5 is the stronger model overall, scoring 47.8 to 41.9 on the Noometry Index. GPT-5.4 nano costs 1.9× less per token, which makes it the better buy when Kimi K2.5's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 nano or Kimi K2.5?
GPT-5.4 nano is cheaper. It lists at $0.20 per million input tokens and $1.25 per million output tokens; Kimi K2.5 lists at $0.45 and $2.25.
Is GPT-5.4 nano or Kimi K2.5 better for coding?
Kimi K2.5 scores higher on coding benchmarks: 48.7 versus 43.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 nano does, with 400K tokens against 262K.
How many benchmarks do GPT-5.4 nano and Kimi K2.5 share?
33 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Kimi K2.5 has 51.