GPT-5.3-Codex vs Kimi K3.1 Pro
Full profiles: GPT-5.3-Codex (OpenAI) and Kimi K3.1 Pro (Moonshot AI): every benchmark, price and rank.
Updated · Data from 9 sources
As of 10 Nov 2026, GPT-5.3-Codex and Kimi K3.1 Pro are neck and neck on quality (62.9 vs 61.2 on the BenchLeader Index, inside the ±7.7 margin), and Kimi K3.1 Pro is 1.2× cheaper at $4.00 vs $4.81 per 1M tokens blended. Choose GPT-5.3-Codex for agents & tools (80 vs 67) and coding (61 vs 60) and fast first answers (39.0 s vs 161.8 s), where it also streams ~2× faster; choose Kimi K3.1 Pro for multimodal (66 vs 61) and long context (68 vs 66).
Key facts (as of 10 Nov 2026)
- GPT-5.3-Codex scores 62.9 on the BenchLeader Index and Kimi K3.1 Pro scores 61.2 (BenchLeader, 10 Nov 2026).
- GPT-5.3-Codex costs $4.81 per 1M tokens blended (list price, 10 Nov 2026).
- Kimi K3.1 Pro costs $4.00 per 1M tokens blended (list price, 10 Nov 2026).
- GPT-5.3-Codex streams ~2× faster than Kimi K3.1 Pro (Artificial Analysis, 7-day median, 10 Nov 2026).
- Of the 6 benchmarks both models have, Kimi K3.1 Pro leads 6 (BenchLeader, 10 Nov 2026).
Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | GPT-5.3-Codex | 61 vs 60 category score | Medium |
| Agents & tool use | GPT-5.3-Codex | 80 vs 67 category score; AA-Briefcase v1.1 868 vs 1674 | High |
| Long documents | Kimi K3.1 Pro | 68 vs 66 category score; AA-LCR v1.1 86.3% vs 83.3% | Medium |
| Images and other media | Kimi K3.1 Pro | 66 vs 61 category score; MMMU-Pro 83.9% vs 78.5% | Medium |
| Lowest cost | Kimi K3.1 Pro | $4.00 vs $4.81 per 1M blended; agentic coding session $0.160 vs $0.161 | – |
| Fastest first answer | GPT-5.3-Codex | 39.0 s (GPT-5.3-Codex (xhigh)) vs 161.8 s (Kimi K3.1 Pro) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- GPT-5.3-Codex (xhigh) and Kimi K3.1 Pro are neck and neck on quality (62.9 vs 61.2, within the ±7.7 margin).
- GPT-5.3-Codex (xhigh) is stronger in agents & tools, coding, instruction following.
- Kimi K3.1 Pro is stronger in long context, multimodal, composite, knowledge, maths, reasoning.
- Kimi K3.1 Pro is 1.2× cheaper ($4.00 vs $4.81 per 1M blended).
- GPT-5.3-Codex (xhigh) streams 2.4× faster (76 vs 32 tokens per second).
At a glance
| Spec | GPT-5.3-Codex | Kimi K3.1 Pro |
|---|---|---|
| Developer | OpenAI | Moonshot AI |
| Released | 5 Feb 2026 | 23 Oct 2026 |
| Context window (tokens) | 400,000 | 1,048,576 |
| Compared at | GPT-5.3-Codex (xhigh) | Kimi K3.1 Pro |
| Open weights | No | Yes |
| Price in / out / cached input per 1M | $1.75 / $14.00 / $0.175 | $2.00 / $10.00 / $0.200 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | GPT-5.3-Codex (xhigh) | Kimi K3.1 Pro | Leader |
|---|---|---|---|
| Reasoning | – | 59.6 | not comparable |
| Coding | 61.0 | 59.8 | GPT-5.3-Codex |
| Agents & tools | 80.1 | 67.1 | GPT-5.3-Codex |
| Maths | – | 73.4 | not comparable |
| Knowledge | – | 61.5 | not comparable |
| Instruction following | 73.6 | – | not comparable |
| Multimodal | 60.9 | 66.4 | Kimi K3.1 Pro |
| Long context | 66.4 | 68.0 | Kimi K3.1 Pro |
| Composite | – | 54.8 | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Reasoning, Maths, Knowledge, Instruction following, Human preference.
Where sources disagree
Across the 6 benchmarks both have, GPT-5.3-Codex leads 0, Kimi K3.1 Pro leads 6.
| Overall index | GPT-5.3-Codex (xhigh) | Kimi K3.1 Pro | Leader |
|---|---|---|---|
| BenchLeader Index | 62.9 | 61.2 | neck and neck |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | GPT-5.3-Codex (xhigh) | Kimi K3.1 Pro |
|---|---|---|
| Blended price per 1M tokens | $4.81 | $4.00 |
| Input / output per 1M | $1.75 / $14.00 | $2.00 / $10.00 |
| Cached input per 1M | $0.175 | $0.200 |
| Price per task: chat reply | $0.0049 | $0.0038 |
| Price per task: summarize a 30-page report | $0.029 | $0.030 |
| Price per task: code edit | $0.032 | $0.027 |
| Price per task: agentic coding session | $0.161 | $0.160 |
| Price per task: structured extraction | $0.0063 | $0.0060 |
| Agentic coding session, input mostly cached | $0.090 | $0.079 |
| Cost per benchmark task (AA Index) | – | $3.68 |
| Output tokens per task (AA Index) | – | 72k |
| Of which reasoning | – | 44k |
| Time per task (AA Index) | – | 38 min |
| Output speed | 76 tok/s | 32 tok/s |
| Time to first answer | 39.0 s | 161.8 s |
| Time to a complete answer | 87.0 s | 397.6 s |
Each model is compared at its best-ranked reasoning-effort setting. GPT-5.3-Codex (xhigh) scores 62.9 and takes 39 s to its first answer token. Kimi K3.1 Pro scores 61.2 and takes 162 s to its first answer token; Kimi K3.1 Pro (max) scores 60.4 in 162 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
86 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | GPT-5.3-Codex (xhigh) | Kimi K3.1 Pro |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 62.9 | 61.2 |
| Agents & tools score | BenchLeader | 80.1 | 67.1 |
| Coding score | BenchLeader | 61.0 | 59.8 |
| Instruction following score | BenchLeader | 73.6 | – |
| Long context score | BenchLeader | 66.4 | 68.0 |
| Multimodal score | BenchLeader | 60.9 | 66.4 |
| Composite score | BenchLeader | – | 54.8 |
| Knowledge score | BenchLeader | – | 61.5 |
| Maths score | BenchLeader | – | 73.4 |
| Reasoning score | BenchLeader | – | 59.6 |
| Blended price $/M | BenchLeader | $4.81 | $4.00 |
| Output speed | BenchLeader | 76 tok/s | 32 tok/s |
| Time to first answer | BenchLeader | 39.0 s | 161.8 s |
| Context window | BenchLeader | 400,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| Arena Hard Prompts | Arena (LMArena) | – | 1501† |
| LiveBench 2.0 Reasoning | LiveBench | – | 91.1% |
| GPQA Diamond (AA) | Artificial Analysis | 91.5% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 42.5% | 52.5% |
| HLE-Diamond (AA) | Artificial Analysis | – | 37.3% |
| CritPt | Artificial Analysis | 16.9% | 24.6% |
| CritPt v2 hard subset (AA) | Artificial Analysis | – | 49.5% |
| MysteryMechanism | Vals AI | – | 26.1%† |
| LMCA | Epoch AI Benchmarking Hub | – | 57.6%† |
| DTBench | Epoch AI Benchmarking Hub | – | 93.3%† |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | 74.8%† | – |
| SciCode | SciCode | – | 60.8% |
| WeirdML | WeirdML | 77.9% | – |
| Arena Coding | Arena (LMArena) | – | 1526† |
| Arena WebDev | Arena (LMArena) | 1406† | 1688† |
| LiveBench 2.0 Coding | LiveBench | – | 67.8% |
| LiveCodeBench | Vals AI | 87.3% | – |
| IOI | Vals AI | 53.8% | 81.9%† |
| SWE-bench (Vals) | Vals AI | 78.0% | – |
| Vibe Code Bench v1.1 | Vals AI | 61.8% | 84.9%† |
| ALE-Bench | Epoch AI Benchmarking Hub | 1655.2 | 1932.8† |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | – | 67.2% |
| Startrise Frontend | Startrise | – | 82.5% |
| GameBench 2 Overall | DuelLab GameBench 2 | – | 78.2 |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | – | 84.0† |
Show all 86 benchmarks
| Metric | Publisher | GPT-5.3-Codex (xhigh) | Kimi K3.1 Pro |
|---|---|---|---|
| Coding | |||
| Design Arena Website | Design Arena | 1159† | 1320 |
| FrontierCode 1.1 (AA) | Artificial Analysis | – | 47.1% |
| FrontierCode 2.0 (AA) | Artificial Analysis | – | 25.4% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 78.4%† | – |
| APEX-Agents | Mercor | – | 65.5% |
| Arena Agent | Arena (LMArena) | – | 5.2† |
| LiveBench Long Agentic | LiveBench | – | 38.7% |
| LiveBench Browser | LiveBench | – | 75.9% |
| Terminal-Bench Hard | Artificial Analysis | 53.0% | – |
| τ²-Bench Telecom (AA) | Artificial Analysis | 86.0% | – |
| HiL-Bench | Scale AI SEAL | 4.3%† | – |
| GDPval-AA v2.1, normalised score | Artificial Analysis | – | 67.3% |
| GDPval-AA v2.1 | Artificial Analysis | – | 1861 |
| Analyst Agent (AA) | Artificial Analysis | – | 38.8% |
| Finance Agent v2.1 | Vals AI | – | 61.5%† |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | – | 11.7%† |
| Legal Research Bench | Vals AI | – | 37.5%† |
| SREBench | Vals AI | – | 27.1%† |
| Tax Agent Bench | Vals AI | – | 24.9%† |
| Terminal-Bench 2.0 (Vals) | Vals AI | 64.0%† | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | – | 63.6%† |
| Terminal-Bench Science | Vals AI | – | 31.4%† |
| METR Time Horizons | METR | 74.5%† | – |
| Vending-Bench 2 | Andon Labs | 5940.1† | – |
| AutomationBench | Artificial Analysis | – | 73.3% |
| GDP.pdf (AA) | Artificial Analysis | – | 31.4% |
| AA-Briefcase v1.1 | Artificial Analysis | 868 | 1674 |
| Design Arena Agentic | Design Arena | 1011† | 1281 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | – | 34.3% |
| CWE-Bench (AA) | Artificial Analysis | – | 60.8% |
| DeepSec-Bench (AA) | Artificial Analysis | – | 36.7% |
| CyberGym end-to-end (AA) | Artificial Analysis | – | 94.7% |
| Maths | |||
| ProofBench | Vals AI | – | 91.0% |
| Knowledge | |||
| LiveBench 2.0 Data Analysis | LiveBench | – | 79.1% |
| LiveBench Research | LiveBench | – | 82.1% |
| AA-Omniscience v1.1 | Artificial Analysis | – | 29.9 |
| BioMysteryBench | Vals AI | – | 75.6%† |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 70.0%† |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | – | 48.9% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | – | 62.7% |
| Instruction following | |||
| IFBench | Artificial Analysis | 75.4% | – |
| Arena Instruction Following | Arena (LMArena) | – | 1485† |
| Human preference | |||
| Arena Text | Arena (LMArena) | – | 1483† |
| Arena Creative Writing | Arena (LMArena) | – | 1457† |
| Arena Multi-turn | Arena (LMArena) | – | 1473† |
| Arena Longer Queries | Arena (LMArena) | – | 1488† |
| Multimodal | |||
| LiveBench Design | LiveBench | – | 73.7% |
| MMMU-Pro | Artificial Analysis | 78.5% | 83.9% |
| AA-3D | Artificial Analysis | – | 1112 |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 83.3% | 86.3% |
| MLCR | Artificial Analysis | – | 38.3% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | – | 162.0 |
| LiveBench 2.0 | LiveBench | – | 68.7% |
| AA Intelligence Index v5.0 | Artificial Analysis | – | 46.5 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 32.5 | 55.1 |
| AA Progress Index v1 | Artificial Analysis | 301.3 | 340.7 |
| Vals Index | Vals AI | – | 62.5† |
| Vals Finance Index | Vals AI | – | 56.0%† |
| Vals Science Index | Vals AI | – | 44.4%† |
| AA Cyber Index | Artificial Analysis | – | 64.1 |
| Metadata | |||
| AA Openness Index | Artificial Analysis | – | 33.3 |
Data as of 2026-11-10. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261103-eaea31df94e6 (https://benchmarks.duellab.org/releases/gb2-20261103-eaea31df94e6/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- Kimi K3.1 Pro: 2 new results (APEX-Agents, Epoch Capabilities Index).
- Kimi K3.1 Pro: BenchLeader Index 61.8 → 61.2.
- Kimi K3.1 Pro: 5 score revisions.
- Kimi K3.1 Pro: BenchLeader Index 61.3 → 61.8.
- Kimi K3.1 Pro: 1 new result (GameBench 2 Overall); 1 score revision.
- Kimi K3.1 Pro: 1 score revision.
- Kimi K3.1 Pro: BenchLeader Index 59.3 → 61.3.
- GPT-5.3-Codex (xhigh): 1 new result (AA-Briefcase v1.1).
- Kimi K3.1 Pro: 3 new results (SciCode, ProofBench, Startrise Frontend).
- Kimi K3.1 Pro: BenchLeader Index 61.8 → 59.3.
Related comparisons
- GPT-5.3-Codex vs Claude Fable 5.2
- GPT-5.3-Codex vs Claude Opus 5.5
- GPT-5.3-Codex vs Gemini 4 Carbon
- Kimi K3.1 Pro vs Claude Fable 5.2
- Kimi K3.1 Pro vs Claude Opus 5.5
- Kimi K3.1 Pro vs Gemini 4 Carbon
- Kimi K3.1 Pro vs GPT-6 Astra
- GPT-5.3-Codex vs Kimi K3
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
GPT-5.3-Codex vs Kimi K3.1 Pro: questions
- Is GPT-5.3-Codex better than Kimi K3.1 Pro?
- GPT-5.3-Codex (xhigh) and Kimi K3.1 Pro are neck and neck on quality (62.9 vs 61.2, within the ±7.7 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 10 Nov 2026, so check the category scores for your use.
- GPT-5.3-Codex vs Kimi K3.1 Pro: which is cheaper, and will that change?
- Kimi K3.1 Pro is cheaper today: $4.00 against $4.81 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is GPT-5.3-Codex or Kimi K3.1 Pro better for coding?
- GPT-5.3-Codex scores higher in coding (61 vs 60 on the category index, where 50 is average).
- Which is faster, GPT-5.3-Codex or Kimi K3.1 Pro?
- GPT-5.3-Codex streams ~2× faster (76 vs 32 output tokens per second, Artificial Analysis 7-day median).
- Is GPT-5.3-Codex or Kimi K3.1 Pro better for agents and tool use?
- GPT-5.3-Codex scores higher in agentic tasks (80 vs 67 on the category index, where 50 is average).
More questions (4)
- Which is better for long documents?
- Kimi K3.1 Pro scores higher on long-context benchmarks (68 vs 66). Kimi K3.1 Pro accepts more context: 1,048,576 against 400,000 tokens for GPT-5.3-Codex.
- Which has lower latency?
- GPT-5.3-Codex answers first: 39.0 s against 161.8 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Do independent benchmarks agree on GPT-5.3-Codex vs Kimi K3.1 Pro?
- Across the 6 benchmarks both have, GPT-5.3-Codex leads 0, Kimi K3.1 Pro leads 6.
- When was Kimi K3.1 Pro released?
- Kimi K3.1 Pro was released on 23 Oct 2026.