Model comparison
Claude Mythos 5 vs Kimi K2.5 (Reasoning)
Head-to-head evidence from 4 shared benchmark results across 3 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: Claude Mythos 5 #1 (Supported); Kimi K2.5 (Reasoning) #57 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. Claude Mythos 5 and Kimi K2.5 (Reasoning) share 4 comparable benchmark results. 4 of 8 categories are comparable. 11 results are unique to Claude Mythos 5; 23 to Kimi K2.5 (Reasoning).
Updated July 23, 2026- Shared results
- 4
- Claude Mythos 5 only
- 11
- Kimi K2.5 (Reasoning) only
- 23
- Comparable categories
- 4 / 8
Pick Claude Mythos 5 if you want the stronger benchmark profile. Kimi K2.5 (Reasoning) only becomes the better choice if knowledge is the priority or you want the cheaper token bill.
Confidence note. This is a partial-evidence comparison with 4 shared benchmark results across 3 evidence categories; 4 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
Claude Mythos 5 is clearly ahead on the BenchAlign aggregate, 83.93 to 59.35. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Claude Mythos 5's sharpest advantage is in agentic, where it averages 87 against 55. The single biggest benchmark swing on the page is Terminal-Bench 2.0, 88% to 50.8%. Kimi K2.5 (Reasoning) does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.
Claude Mythos 5 is also the more expensive model on tokens at $10.00 input / $50.00 output per 1M tokens, versus $0.60 input / $3.00 output per 1M tokens for Kimi K2.5 (Reasoning). That is roughly 16.7x on output cost alone. Claude Mythos 5 gives you the larger context window at 1M+, compared with 128K for Kimi K2.5 (Reasoning).
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 | Claude Mythos 5 | Δ | Kimi K2.5 (Reasoning) |
|---|---|---|---|
| Agentic | Claude Mythos 587.0 | Margin← 32.0 | Kimi K2.5 (Reasoning)55.0 |
| Knowledge | Claude Mythos 568.5 | Margin→ 18.7 | Kimi K2.5 (Reasoning)87.2 |
| Multimodal | Claude Mythos 593.5 | Margin← 15.0 | Kimi K2.5 (Reasoning)78.5 |
| Coding | Claude Mythos 589.7 | Margin← 12.9 | Kimi K2.5 (Reasoning)76.8 |
| Math | Claude Mythos 597.6 | MarginNo overlap | Kimi K2.5 (Reasoning)Not measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
Terminal-Bench 2.0
AgenticA 88%B 50.8%Winner: Claude Mythos 5Δ 37.2Terminal-Bench 2.0: Claude Mythos 5 scored 88%; Kimi K2.5 (Reasoning) scored 50.8%. Claude Mythos 5 wins this benchmark. - Source ↗
BrowseComp
AgenticA 88%B 60.6%Winner: Claude Mythos 5Δ 27.4BrowseComp: Claude Mythos 5 scored 88%; Kimi K2.5 (Reasoning) scored 60.6%. Claude Mythos 5 wins this benchmark. - Source ↗
SWE-bench Verified
CodingA 95.5%B 76.8%Winner: Claude Mythos 5Δ 18.7SWE-bench Verified: Claude Mythos 5 scored 95.5%; Kimi K2.5 (Reasoning) scored 76.8%. Claude Mythos 5 wins this benchmark. - Source ↗
GPQA
KnowledgeA 94.1%B 87.6%Winner: Claude Mythos 5Δ 6.5GPQA: Claude Mythos 5 scored 94.1%; Kimi K2.5 (Reasoning) scored 87.6%. Claude Mythos 5 wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | Claude Mythos 5 | Kimi K2.5 (Reasoning) | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | Claude Mythos 5$10 input / $50 output | Kimi K2.5 (Reasoning)$0.6 input / $3 output | Kimi K2.5 (Reasoning) has the lower combined listed price. |
| Generation speedtokens per second | Claude Mythos 5Not available | Kimi K2.5 (Reasoning)Not available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | Claude Mythos 5Not available | Kimi K2.5 (Reasoning)Not available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | Claude Mythos 51M+ | Kimi K2.5 (Reasoning)128K | Claude Mythos 5 lists the larger context window. |
Benchmark Deep Dive
AgenticClaude Mythos 5 wins10 benchmarks
| Benchmark | Claude Mythos 5 | Kimi K2.5 (Reasoning) | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 88% | 50.8% | Claude Mythos 5 leads |
| OSWorld-VerifiedSource | 85% | — | Not comparable |
| BrowseCompSource | 88% | 60.6% | Claude Mythos 5 leads |
| ExploitGymSource | 17.5% | — | Not comparable |
| APEX-Agents-AASource | — | 11.5% | Not comparable |
| τ²-bench resultsSource | — | 95.9% | Not comparable |
| Gert LabsSource | — | 32.58% | Not comparable |
| AA Agentic IndexSource | — | 21.7% | Not comparable |
| GDPval-AASource | — | 25.4% | Not comparable |
| GDPval-AASource | — | 1009 | Not comparable |
CodingClaude Mythos 5 wins6 benchmarks
| Benchmark | Claude Mythos 5 | Kimi K2.5 (Reasoning) | Result |
|---|---|---|---|
| SWE-bench VerifiedSource | 95.5% | 76.8% | Claude Mythos 5 leads |
| SWE-bench ProSource | 80.3% | — | Not comparable |
| Terminal-Bench 2.0Source | 88.0% | — | Not comparable |
| Vibe Code BenchSource | — | 17.54% | Not comparable |
| AA-SciCodeSource | — | 49.0% | Not comparable |
| AA Coding IndexSource | — | 46.8% | Not comparable |
Reasoning2 benchmarks
KnowledgeKimi K2.5 (Reasoning) wins10 benchmarks
| Benchmark | Claude Mythos 5 | Kimi K2.5 (Reasoning) | Result |
|---|---|---|---|
| GPQASource | 94.1% | 87.6% | Claude Mythos 5 leads |
| HLESource | 64.5% | — | Not comparable |
| HLE w/o toolsSource | 59% | — | Not comparable |
| MMLU-ProSource | — | 87.1% | Not comparable |
| Artificial Analysis Intelligence IndexSource | — | 35.4% | Not comparable |
| AA-GPQA DiamondSource | — | 87.9% | Not comparable |
| AA-HLESource | — | 29.4% | Not comparable |
| AA-Omniscience IndexSource | — | -8.1% | Not comparable |
| AA-Omniscience AccuracySource | — | 34.3% | Not comparable |
| AA-Omniscience Hallucination RateSource | — | 64.6% | Not comparable |
Math2 benchmarks
Multilingual1 benchmarks
| Benchmark | Claude Mythos 5 | Kimi K2.5 (Reasoning) | Result |
|---|---|---|---|
| SWE MultilingualSource | 92.2% | — | Not comparable |
MultimodalClaude Mythos 5 wins6 benchmarks
Inst. Following1 benchmarks
| Benchmark | Claude Mythos 5 | Kimi K2.5 (Reasoning) | Result |
|---|---|---|---|
| AA-IFBenchSource | — | 70.2% | Not comparable |
Frequently Asked Questions (5)
Which is better, Claude Mythos 5 or Kimi K2.5 (Reasoning)?
Claude Mythos 5 is ahead on BenchLM's BenchAlign leaderboard, 83.93 to 59.35. The biggest single separator in this matchup is Terminal-Bench 2.0, where the scores are 88% and 50.8%.
Which is better for knowledge tasks, Claude Mythos 5 or Kimi K2.5 (Reasoning)?
Kimi K2.5 (Reasoning) has the edge for knowledge tasks in this comparison, averaging 87.2 versus 68.5. Inside this category, GPQA is the benchmark that creates the most daylight between them.
Which is better for coding, Claude Mythos 5 or Kimi K2.5 (Reasoning)?
Claude Mythos 5 has the edge for coding in this comparison, averaging 89.7 versus 76.8. Inside this category, SWE-bench Verified is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, Claude Mythos 5 or Kimi K2.5 (Reasoning)?
Claude Mythos 5 has the edge for agentic tasks in this comparison, averaging 87 versus 55. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.
Which is better for multimodal and grounded tasks, Claude Mythos 5 or Kimi K2.5 (Reasoning)?
Claude Mythos 5 has the edge for multimodal and grounded tasks in this comparison, averaging 93.5 versus 78.5. Kimi K2.5 (Reasoning) stays close enough that the answer can still flip depending on your workload.
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.