Model comparison
GPT-5 vs Kimi K2.5
GPT-5 is the stronger model overall, scoring 50.9 to 48.1 on the Noometry Index. Kimi K2.5 costs 3.8× less per token, which makes it the better buy when GPT-5's lead doesn't matter for your workload.
Last verified . 42 shared benchmarks.
Summary
- They share 42 benchmarks with published results for both. GPT-5 scores higher in 6 categories and Kimi K2.5 in 4 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in long context, where GPT-5 leads 69.5 to 52.1.
- The biggest single-benchmark swing is Chess Puzzles: 37% for GPT-5 and 12% for Kimi K2.5.
- Kimi K2.5 is cheaper at $0.45 / $2.25 per million input/output tokens, against $1.25 / $10 for GPT-5.
- GPT-5 accepts more context: 400K tokens versus 262K.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| GPT-5 | Kimi K2.5 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 50.9 | 48.1 |
| Released | 2025-08-07 | 2026-01-27 |
| Weights | Proprietary | Open |
| Context window | 400K | 262K |
| Max output | 128K | 262K |
| Input $ / M tokens | $1.25 | $0.45 |
| Output $ / M tokens | $10 | $2.25 |
| Results tracked | 69 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
GPT-5: 50.3 (#47), Kimi K2.5: 48.8 (#53)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| SWE-bench Verified | 73.6% | 73.8% |
| SWE-bench Verified (bash only) | 65% | 70.8% |
| LMArena WebDev | 1418 | 1437 |
| SciCode | 42.9% | 49% |
| WeirdML | 60.7% | 45.6% |
| LMArena Coding | 1436 | 1474 |
| ALE-Bench | 1,162 | 821.65 |
| Aider Polyglot | 88% | — |
| SWE-bench Multilingual | — | 67.3% |
| GSO | 6.9% | — |
| AlgoTune | 1.67 | — |
Agentic & Tool Use Kimi K2.5 leads
GPT-5: 33.1 (#56), Kimi K2.5: 34.2 (#48)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | 49.6% | 43.2% |
| GDPval | 34.8% | — |
| Remote Labor Index | 1.7% | — |
| DeepResearch Bench | 49.6% | — |
| OSWorld | — | 63.3% |
| BALROG | 32.8% | — |
| LMArena Search | 1133 | — |
| METR Time Horizons | 69.6% | — |
| Vending-Bench 2 | — | 1,198 |
Reasoning GPT-5 leads
GPT-5: 38.3 (#64), Kimi K2.5: 31.2 (#80)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 9.9% | 11.8% |
| SimpleBench | 56.7% | 46.8% |
| Kagi LLM Benchmark | 72.7% | 78.5% |
| ARC-AGI-1 | 65.7% | 65.3% |
| CritPt | 12.6% | 3.1% |
| Chess Puzzles | 37% | 12% |
| EnigmaEval | 10.5% | 3.4% |
| LMArena Hard Prompts | 1416 | 1453 |
| Epoch Capabilities Index | 150 | 148.03 |
| NYT Connections (extended) | — | 69.9% |
| Thematic Generalization | — | 69.4% |
| EBR-Bench | 12.7% | — |
| Mystery Game Puzzles | 23% | — |
| DTBench | 90.7% | — |
| LMCA | 40% | — |
| ForecastBench | 61.4 | — |
Math GPT-5 leads
GPT-5: 55.0 (#44), Kimi K2.5: 51.8 (#53)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 91.4% | 92.2% |
| LMArena Math | 1407 | 1470 |
| FrontierMath (Feb 2025 set) | 32.4% | 27.9% |
| FrontierMath Tier 4 (v1) | 12.5% | 4.2% |
| FrontierMath (Tiers 1-3) | 55.4% | — |
| FrontierMath Tier 4 | 22% | — |
| MathArena Final-Answer Competitions | — | 62.3% |
| ProofBench | 18% | — |
| Omni-MATH | 64.7% | — |
| MATH Level 5 | 98.1% | — |
Knowledge GPT-5 leads
GPT-5: 56.6 (#43), Kimi K2.5: 53.6 (#56)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 86.2% | 87.6% |
| Humanity's Last Exam | 25.3% | 24.4% |
| SimpleQA Verified | 50.1% | 34.3% |
| Vectara Hallucination Rate | 14.7% | 14.2% |
| LMArena Expert | 1419 | 1466 |
| MMLU-Pro | 86.3% | — |
| Confabulations | 10.3% | — |
| GPQA (HELM) | 79.2% | — |
Multimodal GPT-5 leads
GPT-5: 46.8 (#13), Kimi K2.5: 41.1 (#39)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1232 | 1269 |
| GeoBench | 81% | — |
| VPCT | 66% | — |
| LMArena Document | — | 1430 |
Multilingual Kimi K2.5 leads
GPT-5: 51.4 (#110), Kimi K2.5: 53.9 (#53)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1397 | 1433 |
| LMArena Chinese | 1422 | 1495 |
| LMArena French | 1410 | 1454 |
| LMArena German | 1416 | 1441 |
| LMArena Japanese | 1409 | 1421 |
| LMArena Korean | 1360 | 1410 |
| LMArena Russian | 1406 | 1435 |
| LMArena Spanish | 1399 | 1450 |
Instruction Following Kimi K2.5 leads
GPT-5: 73.8 (#113), Kimi K2.5: 75.3 (#64)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1388 | 1431 |
| IFEval | 87.5% | — |
Long Context GPT-5 leads
GPT-5: 69.5 (#2), Kimi K2.5: 52.1 (#7)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| Fiction.LiveBench | 97.2% | 86.1% |
| LMArena Longer Query | 1399 | 1445 |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Kimi K2.5 leads
GPT-5: 63.4 (#65), Kimi K2.5: 65.1 (#53)
| Benchmark | GPT-5 | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1406 | 1445 |
| LMArena Creative Writing | 1365 | 1423 |
| EQ-Bench Creative Writing | 1627 | 1579 |
| LMArena Multi-Turn | 1426 | 1444 |
| Short-Story Creative Writing | 86% | — |
| WildBench | 85.7% | — |
Frequently asked questions
Is GPT-5 better than Kimi K2.5?
GPT-5 is the stronger model overall, scoring 50.9 to 48.1 on the Noometry Index. Kimi K2.5 costs 3.8× less per token, which makes it the better buy when GPT-5's lead doesn't matter for your workload.
Which is cheaper, GPT-5 or Kimi K2.5?
Kimi K2.5 is cheaper. It lists at $0.45 per million input tokens and $2.25 per million output tokens; GPT-5 lists at $1.25 and $10.
Is GPT-5 or Kimi K2.5 better for coding?
GPT-5 scores higher on coding benchmarks: 50.3 versus 48.8 in the Noometry coding category.
Which has the bigger context window?
GPT-5 does, with 400K tokens against 262K.
How many benchmarks do GPT-5 and Kimi K2.5 share?
42 benchmarks have published results for both models. GPT-5 has 69 scored results on Noometry and Kimi K2.5 has 51.