Model comparison
Kimi K2.5 vs o3-mini
Head-to-head evidence from 8 shared benchmark results across 4 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: Kimi K2.5 #54 (Supported); o3-mini #136 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. Kimi K2.5 and o3-mini share 8 comparable benchmark results. 3 of 8 categories are comparable. 55 results are unique to Kimi K2.5; 2 to o3-mini.
Updated July 23, 2026- Shared results
- 8
- Kimi K2.5 only
- 55
- o3-mini only
- 2
- Comparable categories
- 3 / 8
Pick Kimi K2.5 if you want the stronger benchmark profile. o3-mini only becomes the better choice if knowledge is the priority or you want the stronger reasoning-first profile.
Confidence note. This is a partial-evidence comparison with 8 shared benchmark results across 4 evidence categories; 3 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
Kimi K2.5 is clearly ahead on the BenchAlign aggregate, 59.66 to 47.41. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Kimi K2.5's sharpest advantage is in coding, where it averages 59.4 against 49.3. The single biggest benchmark swing on the page is SWE-bench Verified, 76.8% to 49.3%. o3-mini does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.
o3-mini is also the more expensive model on tokens at $1.10 input / $4.40 output per 1M tokens, versus $0.60 input / $3.00 output per 1M tokens for Kimi K2.5. o3-mini is the reasoning model in the pair, while Kimi K2.5 is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. Kimi K2.5 gives you the larger context window at 256K, compared with 200K for o3-mini.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | Kimi K2.5 | Δ | o3-mini |
|---|---|---|---|
| Knowledge | Kimi K2.556.9 | Margin→ 20.3 | o3-mini77.2 |
| Coding | Kimi K2.559.4 | Margin← 10.1 | o3-mini49.3 |
| Inst. Following | Kimi K2.593.9 | MarginTie | o3-mini93.9 |
| Agentic | Kimi K2.555.0 | MarginNo overlap | o3-miniNot measured |
| Reasoning | Kimi K2.561.0 | MarginNo overlap | o3-miniNot measured |
| Math | Kimi K2.560.6 | MarginNo overlap | o3-miniNot measured |
| Multilingual | Kimi K2.582.3 | MarginNo overlap | o3-miniNot measured |
| Multimodal | Kimi K2.578.5 | MarginNo overlap | o3-miniNot measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
SWE-bench Verified
CodingA 76.8%B 49.3%Winner: Kimi K2.5Δ 27.5SWE-bench Verified: Kimi K2.5 scored 76.8%; o3-mini scored 49.3%. Kimi K2.5 wins this benchmark. - Source ↗
GPQA
KnowledgeA 87.6%B 77.2%Winner: Kimi K2.5Δ 10.4GPQA: Kimi K2.5 scored 87.6%; o3-mini scored 77.2%. Kimi K2.5 wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | Kimi K2.5 | o3-mini | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | Kimi K2.5$0.6 input / $3 output | o3-mini$1.1 input / $4.4 output | Kimi K2.5 has the lower combined listed price. |
| Generation speedtokens per second | Kimi K2.545 tok/s | o3-mini160 tok/s | o3-mini has the higher measured throughput. |
| First-answer latencyseconds to first token | Kimi K2.52.38 s | o3-mini7.12 s | Kimi K2.5 reaches the first token sooner. |
| Context windowmaximum listed tokens | Kimi K2.5256K | o3-mini200K | Kimi K2.5 lists the larger context window. |
Benchmark Deep Dive
Agentic19 benchmarks
| Benchmark | Kimi K2.5 | o3-mini | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 50.8% | — | Not comparable |
| BrowseCompSource | 60.6% | — | Not comparable |
| Claw-EvalSource | 52.3% | — | Not comparable |
| QwenClawBenchSource | 54.3% | — | Not comparable |
| τ³-bench resultsSource | 65.7% | — | Not comparable |
| DeepSearchQASource | 77.1% | — | Not comparable |
| DeepPlanningSource | 14.4% | — | Not comparable |
| ToolathlonSource | 27.8% | — | Not comparable |
| MCP AtlasSource | 29.5% | — | Not comparable |
| MCP-TasksSource | 59.1% | — | Not comparable |
| WideResearchSource | 72.7% | — | Not comparable |
| τ²-bench resultsSource | 95.9% | 28.7% | Kimi K2.5 leads |
| APEX-Agents-AASource | 11.5% | — | Not comparable |
| Gert LabsSource | 45.88% | — | Not comparable |
| ResearchClawBenchSource | 14.0% | — | Not comparable |
| JobBenchSource | 8.7% | — | Not comparable |
| AA Agentic IndexSource | 21.7% | — | Not comparable |
| GDPval-AASource | 25.4% | — | Not comparable |
| GDPval-AASource | 1009 | — | Not comparable |
CodingKimi K2.5 wins10 benchmarks
| Benchmark | Kimi K2.5 | o3-mini | Result |
|---|---|---|---|
| SWE-bench VerifiedSource | 76.8% | 49.3% | Kimi K2.5 leads |
| SWE-bench Verified*Source | 70.8% | — | Not comparable |
| LiveCodeBench v6Source | 85.0% | — | Not comparable |
| SWE-bench ProSource | 50.7% | — | Not comparable |
| SWE MultilingualSource | 73% | — | Not comparable |
| SWE-RebenchSource | 58.5% | — | Not comparable |
| React Native EvalsSource | 77.2% | — | Not comparable |
| SciCodeSource | 48.7% | — | Not comparable |
| AA-SciCodeSource | 49.0% | 39.9% | Kimi K2.5 leads |
| AA Coding IndexSource | 46.8% | — | Not comparable |
Reasoning3 benchmarks
Knowledgeo3-mini wins13 benchmarks
| Benchmark | Kimi K2.5 | o3-mini | Result |
|---|---|---|---|
| GPQASource | 87.6% | 77.2% | Kimi K2.5 leads |
| GPQA-DSource | 87.6% | — | Not comparable |
| SuperGPQASource | 69.2% | — | Not comparable |
| MMLU-ProSource | 87.1% | — | Not comparable |
| MMLU-Pro (Arcee)Source | 87.1% | — | Not comparable |
| HLESource | 30.1% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 35.4% | 19.0% | Kimi K2.5 leads |
| AA-GPQA DiamondSource | 87.9% | 74.8% | Kimi K2.5 leads |
| AA-HLESource | 29.4% | 8.7% | Kimi K2.5 leads |
| AA-Omniscience IndexSource | -8.1% | — | Not comparable |
| AA-Omniscience AccuracySource | 34.3% | — | Not comparable |
| AA-Omniscience Hallucination RateSource | 64.6% | — | Not comparable |
| MMLUSource | — | 86.9% | Not comparable |
Math10 benchmarks
| Benchmark | Kimi K2.5 | o3-mini | Result |
|---|---|---|---|
| AIME 2025Source | 96.1% | — | Not comparable |
| AIME26Source | 95.8% | — | Not comparable |
| AIME25 (Arcee)Source | 96.3% | — | Not comparable |
| HMMT Feb 2025Source | 95.4% | — | Not comparable |
| HMMT Nov 2025Source | 91.1% | — | Not comparable |
| HMMT Feb 2026Source | 87.1% | — | Not comparable |
| MMAnswerBenchSource | 81.8% | — | Not comparable |
| FrontierMath v2 (Tiers 1-3)Source | 27.900% | — | Not comparable |
| FrontierMath v2 (Tier 4)Source | 4.200% | — | Not comparable |
| AIME 2024Source | — | 87.3% | Not comparable |
Multilingual2 benchmarks
Multimodal6 benchmarks
Frequently Asked Questions (4)
Which is better, Kimi K2.5 or o3-mini?
Kimi K2.5 is ahead on BenchLM's BenchAlign leaderboard, 59.66 to 47.41. The biggest single separator in this matchup is SWE-bench Verified, where the scores are 76.8% and 49.3%.
Which is better for knowledge tasks, Kimi K2.5 or o3-mini?
o3-mini has the edge for knowledge tasks in this comparison, averaging 77.2 versus 56.9. Inside this category, AA-HLE is the benchmark that creates the most daylight between them.
Which is better for coding, Kimi K2.5 or o3-mini?
Kimi K2.5 has the edge for coding in this comparison, averaging 59.4 versus 49.3. Inside this category, SWE-bench Verified is the benchmark that creates the most daylight between them.
Which is better for instruction following, Kimi K2.5 or o3-mini?
Kimi K2.5 and o3-mini are effectively tied for instruction following here, both landing at 93.9 on average.
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Related Comparisons
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.