Across 18 comparable benchmarks, Kimi K2.6 has better reported results on 15 and worse on 3.
Moderate evidence18 of 18 shared results countedHow we compare
Price basis: Standard rates vs international rates. The price section below shows the applicable periods.
| Fact | GPT-5 | Kimi K2.6 |
|---|---|---|
| Context window | 400K tokens | 256K tokens |
| Input / 1M | $1.25 | $0.950 |
| Output / 1M | $10.00 | $4.00 |
| Blended (4:1) | $2.19 | $0.991 |
| Status | Deprecated | Available |
| Released | August 7, 2025 | April 20, 2026 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Mean of 13 shared benchmarks grouped by capability on a 0–100 scale. How averages are built
Kimi K2.6 ahead on 4 of 4 counted
Kimi K2.6 ahead on 2 of 2 counted
Split 2–2 of 4 counted
Kimi K2.6 ahead on 1 of 1 counted
GPT-5 ahead on 1 of 1 counted
Kimi K2.6 ahead on 1 of 1 counted
Kimi K2.6 ahead on 3 of 3 counted
No shared comparable results for Abstract reasoning, Computer & tool use, Cybersecurity.
Published scores, with matching modes where available. Gaps show GPT-5 relative to Kimi K2.6. Reading the results
| Benchmark | GPT-5 | Kimi K2.6 | Gap |
|---|---|---|---|
| SWE-Bench Pro - PublicRepository Engineering · Accuracy | 41.8Thinking High · tools | 58.6≠Thinking · tools | −16.8 pts#59 vs #22 of 66 |
| SWE-bench VerifiedRepository Engineering · Accuracy | 72.8Thinking High · no tools | 80.2≠Thinking · tools | −7.4 pts#49 vs #15 of 108 |
| Terminal Bench HardAgentic Development · Accuracy | 37.9Thinking Medium · tools | 43.9≠Thinking · tools | −6.0 pts#38 vs #23 of 157 |
| Vibe Code Bench v1.1Code Generation & Editing · Pass rate (%) | 20.1Thinking High · tools | 37.9≠Unspecified Reasoning · tools | −17.8 pts#88 vs #74 of 121 |
| FrontierMath Tier 4 v2Mathematics · Accuracy (verification_code) | 22.0Thinking High · tools | 25.6≠Reported best | −3.7 pts#49 vs #46 of 73 |
| FrontierMath v2Mathematics · Accuracy (verification_code) | 55.4Thinking High · tools | 57.2≠Reported best | −1.8 pts#52 vs #48 of 85 |
| CritPtScientific Reasoning · Score | 5.7Thinking High · no tools | 8.0≠Thinking · no tools | 0.71×#66 vs #62 of 134 |
| GPQA DiamondScientific Reasoning · Accuracy | 87.3Thinking · tools | 90.5≠Thinking · no tools | −3.2 pts#69 vs #46 of 200 |
| MedCodeClinical Workflows · Accuracy (%) | 49.6Thinking High · tools | 40.1≠Unspecified Reasoning · tools | +9.5 pts#21 vs #63 of 97 |
| MedScribeClinical Workflows · Rubric pass rate (%) | 83.7Thinking High · tools | 78.2≠Unspecified Reasoning · tools | +5.5 pts#42 vs #61 of 99 |
| BrowseCompFact Finding · Accuracy | 54.9Thinking · tools | 83.2≠Thinking · tools + web | −28.3 pts#45 vs #20 of 61 |
| EBR-benchMemory & Persistence · Score (% of 21 objectives, best of the two scored playthroughs) | 12.7Thinking High · tools | 2.4≠Thinking · tools | +10.3 pts#23 vs #29 of 30 |
| MMMU-ProVisual Understanding · Accuracy | 78.4Thinking · no tools | 79.4Thinking · no tools | −1.0 pts#45 vs #36 of 136 |
| Creative WritingWriting | 1626Normal · no tools | 1721Normal · no tools | 0.95×#42 vs #30 of 121 |
| SAGEService Workflows · Accuracy (%) | 43.7Thinking High · tools | 50.2≠Unspecified Reasoning · tools | −6.5 pts#47 vs #17 of 82 |
| τ³-BankingService Workflows · Score | 22.1Thinking High · tools | 23.3≠Thinking · tools | 0.95×#55 vs #51 of 120 |
| HLEKnowledge Exams · Accuracy | 35.2Thinking · tools | 54.0≠Thinking · tools + web | −18.8 pts#82 vs #33 of 151 |
| ECICapability Indices · ECI score (capability index, higher is better) | 150Reported best | 151Reported best | 0.99×#49 vs #45 of 168 |
Repository Engineering · Accuracy
Repository Engineering · Accuracy
Agentic Development · Accuracy
Code Generation & Editing · Pass rate (%)
Mathematics · Accuracy (verification_code)
Mathematics · Accuracy (verification_code)
Scientific Reasoning · Score
Scientific Reasoning · Accuracy
Clinical Workflows · Accuracy (%)
Clinical Workflows · Rubric pass rate (%)
Fact Finding · Accuracy
Memory & Persistence · Score (% of 21 objectives, best of the two scored playthroughs)
Visual Understanding · Accuracy
Writing
Service Workflows · Accuracy (%)
Service Workflows · Score
A dash means no result is available.
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix · 5% steps
Blended price = effective input rate at 90% cache hits × 80% + output rate × 20%
GPT-5 costs121%morethan Kimi K2.6 per million tokens — 2.2× the price.Using standard rates vs international rates at this token mix.
GPT-5: Standard pricing. OpenAI official pricing (developers.openai.com/api/docs/pricing) verified 2026-11-13; standard tier; Deprecated; shutdown 2026-12-11 (replacement: gpt-6.1-sol)
Kimi K2.6: International pricing. Kimi Intl standard
| Price / 1M tokens | GPT-5 | Kimi K2.6 | Ratio |
|---|---|---|---|
| Input | $1.25 | $0.95 | 1.32× |
| Output | $10.00 | $4.00 | 2.50× |
| Cached input | $0.125 | $0.16 | 0.78× |
| Cache write | — | — | — |
| Blended | $2.19 | $0.991 | 2.21× |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
Reported cost per task, paired with the score from the same run. Settings and sources may differ; costs do not affect the verdict.
| Benchmark | GPT-5 | Kimi K2.6 |
|---|---|---|
| Agents' Last Exam v1.0 | Not reported | |
| AIME 2026 | Not reported | |
| ARC-AGI-1 | 1 more settings | Not reported |
| ARC-AGI-2 | 1 more settings$0.4486 / task Average · ARC-AGI-2 Score: 7.5 Thinking Medium (No Tools) | Not reported |
| AutomationBench | Not reported | $0.28 / task Average · Pass Rate: 2.6 Reported best (effort unspecified) · zapier-private-heldout |
| Code Migration | Not reported | $5.1121 / task Average · Hidden-test pass rate (%): 27.8 Unspecified Reasoning (With Tools) |
| EMB (Excel Modeling) | Not reported | |
| Finance Agent v2 | Not reported | |
| Legal Research Bench | Not reported | |
| MedCode | Not reported | |
| MedScribe | Not reported | |
| Public Benefits Bench v1.1 | Not reported | $0.9063 / task Average · Criterion pass rate (%): 56.6 Unspecified Reasoning (With Tools) |
| SAGE | Not reported | |
| Tax Agent Bench | Not reported | $0.5655 / task Average · Accuracy (%, rubric × citation quality): 12.4 Unspecified Reasoning (With Tools) |
| Vibe Code Bench v1.1 | Not reported |
| Attribute | GPT-5 | Kimi K2.6 |
|---|---|---|
| Availability | Deprecated | Available |
| Provider | OpenAI | Moonshot AI |
| Model type | Foundation model | Reasoning model |
| First release | August 7, 2025 | April 20, 2026 |
| Context window | 400K tokens | 256K tokens |
| Max output | 128K tokens | — |
| Total parameters | — | 1.0T |
| Active parameters | — | 32B |
| Architecture | Undisclosed | Mixture of Experts (MoE) |
| Documented capabilities | Reasoning, Multilingual | Reasoning, Multilingual, Coding, Tool use, Function calling, Computer use, Long context |
| Knowledge cutoff | — | 2025-04 |
| Input modalities | Text, Image | Text, Image, Video |
| Output modalities | Text | Text |
| Weights | Closed / API only | Open weights |
| Licence | — | Modified MIT |
| Commercial use | Proprietary | Commercial use allowed |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used GPT-5 or Kimi K2.6? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.