Model comparison
GPT-5.6 Terra vs Kimi K3
Head-to-head evidence from 27 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: GPT-5.6 Terra #11 (Estimated); Kimi K3 #4 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.6 Terra and Kimi K3 share 27 comparable benchmark results. 3 of 8 categories are comparable. 17 results are unique to GPT-5.6 Terra; 30 to Kimi K3.
Updated July 23, 2026- Shared results
- 27
- GPT-5.6 Terra only
- 17
- Kimi K3 only
- 30
- Comparable categories
- 3 / 8
Pick Kimi K3 if you want the stronger benchmark profile. GPT-5.6 Terra only becomes the better choice if knowledge is the priority.
Confidence note. This is a partial-evidence comparison with 27 shared benchmark results across 5 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 K3 is clearly ahead on the BenchAlign aggregate, 80.96 to 72.57. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Kimi K3's sharpest advantage is in agentic, where it averages 89.5 against 87.4. The single biggest benchmark swing on the page is BrowseComp, 87.5% to 91.2%. GPT-5.6 Terra does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.
Kimi K3 gives you the larger context window at 1.05M, compared with 1M for GPT-5.6 Terra.
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 | GPT-5.6 Terra | Δ | Kimi K3 |
|---|---|---|---|
| Knowledge | GPT-5.6 Terra92.9 | Margin← 31.9 | Kimi K361.0 |
| Multimodal | GPT-5.6 Terra80.7 | Margin← 2.2 | Kimi K378.5 |
| Agentic | GPT-5.6 Terra87.4 | Margin→ 2.1 | Kimi K389.5 |
| Coding | GPT-5.6 Terra63.4 | MarginNo overlap | Kimi K3Not measured |
| Math | GPT-5.6 Terra80.8 | MarginNo overlap | Kimi K3Not measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
BrowseComp
AgenticA 87.5%B 91.2%Winner: Kimi K3Δ 3.7BrowseComp: GPT-5.6 Terra scored 87.5%; Kimi K3 scored 91.2%. Kimi K3 wins this benchmark. - Source ↗
Terminal-Bench 2.0
AgenticA 87.4%B 88.3%Winner: Kimi K3Δ 0.9Terminal-Bench 2.0: GPT-5.6 Terra scored 87.4%; Kimi K3 scored 88.3%. Kimi K3 wins this benchmark. - Source ↗
MMMU-Pro
MultimodalA 80.7%B 81.6%Winner: Kimi K3Δ 0.9MMMU-Pro: GPT-5.6 Terra scored 80.7%; Kimi K3 scored 81.6%. Kimi K3 wins this benchmark. - Source ↗
GPQA
KnowledgeA 92.9%B 93.5%Winner: Kimi K3Δ 0.6GPQA: GPT-5.6 Terra scored 92.9%; Kimi K3 scored 93.5%. Kimi K3 wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.6 Terra | Kimi K3 | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.6 Terra$2.5 input / $15 output | Kimi K3$3 input / $15 output | GPT-5.6 Terra has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.6 TerraNot available | Kimi K3Not available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | GPT-5.6 TerraNot available | Kimi K3Not available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | GPT-5.6 Terra1M | Kimi K31.05M | Kimi K3 lists the larger context window. |
Benchmark Deep Dive
AgenticKimi K3 wins27 benchmarks
| Benchmark | GPT-5.6 Terra | Kimi K3 | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 87.4% | 88.3% | Kimi K3 leads |
| BrowseCompSource | 87.5% | 91.2% | Kimi K3 leads |
| OSWorld 2.0Source | 50.2% | — | Not comparable |
| CyberGymSource | 81.8% | — | Not comparable |
| ExploitGymSource | 23.2% | — | Not comparable |
| ToolathlonSource | 53.1% | — | Not comparable |
| AA Agentic IndexSource | 47.4% | 50.1% | Kimi K3 leads |
| τ²-bench resultsSource | 86.3% | — | Not comparable |
| GDPval-AASource | 54.1% | 59.0% | Kimi K3 leads |
| GDPval-AASource | 1581 | 1679 | Kimi K3 leads |
| AA Harvey LABSource | 85.2% | 94.6% | Kimi K3 leads |
| AA ITBenchSource | 51.0% | 47.7% | GPT-5.6 Terra leads |
| AA Tau3 BankingSource | 31.8% | 33.4% | Kimi K3 leads |
| AA AutomationBenchSource | 45.6% | 52.7% | Kimi K3 leads |
| terminalBenchHardSource | 57.6% | — | Not comparable |
| aaTerminalBench21Source | 88% | 85% | GPT-5.6 Terra leads |
| APEX-Agents-AASource | 38.9% | 41.3% | Kimi K3 leads |
| DeepSearchQASource | — | 95.0% | Not comparable |
| Toolathlon-VerifiedSource | — | 73.2% | Not comparable |
| MCP AtlasSource | — | 84.2% | Not comparable |
| AutomationBenchSource | — | 30.8% | Not comparable |
| JobBenchSource | — | 52.9% | Not comparable |
| APEX-AgentsSource | — | 37.6% | Not comparable |
| SpreadsheetBench 2Source | — | 34.8% | Not comparable |
| DECK-BenchSource | — | 73.5% | Not comparable |
| AA BriefcaseSource | — | 1543 | Not comparable |
| AA EnterpriseOps-GymSource | — | 45.3% | Not comparable |
Coding13 benchmarks
| Benchmark | GPT-5.6 Terra | Kimi K3 | Result |
|---|---|---|---|
| SWE-bench ProSource | 63.4% | — | Not comparable |
| Terminal-Bench 2.0Source | 87.4% | — | Not comparable |
| deepSweSource | 69.6% | 67.5% | GPT-5.6 Terra leads |
| FrontierCode 1.1 ExtendedSource | 55.8% | — | Not comparable |
| cursorBench32Source | 64.9% | — | Not comparable |
| AA Coding IndexSource | 76.7% | 76.2% | GPT-5.6 Terra leads |
| AA-SciCodeSource | 53.9% | 58.7% | Kimi K3 leads |
| FrontierSWESource | — | 81.2% | Not comparable |
| ProgramBenchSource | — | 77.8% | Not comparable |
| Kimi Code Bench v2Source | — | 72.9% | Not comparable |
| sweMarathonSource | — | 42% | Not comparable |
| PostTrain BenchSource | — | 36.6% | Not comparable |
| MLS-Bench LiteSource | — | 48.3% | Not comparable |
Reasoning3 benchmarks
KnowledgeGPT-5.6 Terra wins12 benchmarks
| Benchmark | GPT-5.6 Terra | Kimi K3 | Result |
|---|---|---|---|
| GPQASource | 92.9% | 93.5% | Kimi K3 leads |
| GPQA-DSource | 92.9% | 93.5% | Kimi K3 leads |
| HealthBench ProfessionalSource | 57.7% | — | Not comparable |
| HealthBench HardSource | 32.7% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 55.0% | 57.1% | Kimi K3 leads |
| AA-GPQA DiamondSource | 92.5% | 93.5% | Kimi K3 leads |
| AA-HLESource | 41.8% | 44.3% | Kimi K3 leads |
| AA-Omniscience IndexSource | -0.2% | 18.4% | Kimi K3 leads |
| AA-Omniscience AccuracySource | 45.9% | 46.0% | Kimi K3 leads |
| AA-Omniscience Hallucination RateSource | 85.2% | 50.9% | Kimi K3 leads |
| HLESource | — | 56% | Not comparable |
| HLE w/o toolsSource | — | 43.5% | Not comparable |
Math3 benchmarks
MultimodalGPT-5.6 Terra wins15 benchmarks
| Benchmark | GPT-5.6 Terra | Kimi K3 | Result |
|---|---|---|---|
| MMMU-ProSource | 80.7% | 81.6% | Kimi K3 leads |
| MMMU-Pro w/ PythonSource | 82% | 83.4% | Kimi K3 leads |
| AA-MMMU-ProSource | 80.7% | 80.5% | GPT-5.6 Terra leads |
| OfficeQA ProSource | — | 63.3% | Not comparable |
| CharXiv w/o toolsSource | — | 84.8% | Not comparable |
| CharXivSource | — | 91.3% | Not comparable |
| MathVisionSource | — | 94.3% | Not comparable |
| MathVision w/ PythonSource | — | 97.8% | Not comparable |
| BabyVision w/ PythonSource | — | 85.7% | Not comparable |
| ZeroBenchSource | — | 23.0% | Not comparable |
| ZeroBench w/ PythonSource | — | 41.0% | Not comparable |
| WorldVQA ForceAnswerSource | — | 51.0% | Not comparable |
| OmniDocBenchSource | — | 91.1% | Not comparable |
| PerceptionBenchSource | — | 58.5% | Not comparable |
| Design Arena WebsiteSource | — | 1386 | Not comparable |
Inst. Following1 benchmarks
| Benchmark | GPT-5.6 Terra | Kimi K3 | Result |
|---|---|---|---|
| AA-IFBenchSource | 71.2% | — | Not comparable |
Frequently Asked Questions (4)
Which is better, GPT-5.6 Terra or Kimi K3?
Kimi K3 is ahead on BenchLM's BenchAlign leaderboard, 80.96 to 72.57. The biggest single separator in this matchup is BrowseComp, where the scores are 87.5% and 91.2%.
Which is better for knowledge tasks, GPT-5.6 Terra or Kimi K3?
GPT-5.6 Terra has the edge for knowledge tasks in this comparison, averaging 92.9 versus 61. Inside this category, AA-Omniscience Hallucination Rate is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, GPT-5.6 Terra or Kimi K3?
Kimi K3 has the edge for agentic tasks in this comparison, averaging 89.5 versus 87.4. Inside this category, GDPval-AA is the benchmark that creates the most daylight between them.
Which is better for multimodal and grounded tasks, GPT-5.6 Terra or Kimi K3?
GPT-5.6 Terra has the edge for multimodal and grounded tasks in this comparison, averaging 80.7 versus 78.5. Inside this category, MMMU-Pro w/ Python is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
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.