Kimi K3.1 Pro (Together, max): benchmark scores, price and ranking
Kimi K3.1 Pro (Together, max) from Moonshot ranks #23 of 827 LLMs on the LLMs Tiger aggregate benchmark score (55.3), based on 8 benchmark results captured up to 2026-11-18.
- Organisation
- Moonshot
- Type
- Open-weight / local model
- Open weights
- Yes
- Released
- 2026-10-23 (source: epoch)
- Cheapest API price
- $0.240 in / $8.40 out per million tokens (via openrouter, observed 2026-11-18T15:47Z)
- Aggregate rank
- #23 of 827 (score 55.3)
Relative to other models, Kimi K3.1 Pro (Together, max) is strongest on GPQA Diamond (#12 of 302) and weakest on Furniture Assembly (#14 of 36).
Across the 8 benchmarks where at least ten models are scored, Kimi K3.1 Pro (Together, max) is in the top 10% on 2 and in the top half on 8.
Its cheapest listed input price, $0.240 per million tokens, is lower than that of 67% of the 443 models with a listed price.
Kimi K3.1 Pro (Together, max) benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| GPQA Diamond | 94.2% | #12 of 302 | Epoch AI Benchmarking Hub | 2026-11-18 |
| OTIS Mock AIME | 100.0% | #12 of 206 | Epoch AI Benchmarking Hub | 2026-11-18 |
| SimpleQA Verified | 62.2% | #13 of 83 | Epoch AI Benchmarking Hub | 2026-11-18 |
| FrontierMath T1-3 | 78.2% | #21 of 78 | Epoch AI Benchmarking Hub | 2026-11-18 |
| Mystery Games | 42.0% | #21 of 69 | Epoch AI Benchmarking Hub | 2026-11-18 |
| Chess Puzzles | 45.0% | #26 of 84 | Epoch AI Benchmarking Hub | 2026-11-18 |
| FrontierMath T4 | 65.9% | #19 of 60 | Epoch AI Benchmarking Hub | 2026-11-18 |
| Furniture Assembly | 55.8% | #14 of 36 | Epoch AI Benchmarking Hub | 2026-11-18 |
About Kimi K3.1 Pro (Together, max)
Kimi K3.1 Pro (Together, max) is an open-weight model that can run on your own hardware from Moonshot. It was released on 2026-10-23, according to Epoch AI. The cheapest API listing we track charges $0.240 per million input tokens and $8.40 per million output tokens, through openrouter. We hold 8 benchmark results for it across 2 areas, and 2 are in the top tenth of their boards.
Where Kimi K3.1 Pro (Together, max) stands
On our aggregate score, Kimi K3.1 Pro (Together, max) is #23 of 827 ranked models, in the top fifty. Its score is 55.3; the scale is centred on 50, so it sits 5.3 points above the typical model. That beats 97% of the ranked field. Among the 15 ranked Moonshot models it is #2. The next one up is Kimi K3.1 Pro (max) at 55.5. The next one down is Kimi K3 (max) at 55.2.
Kimi K3.1 Pro (Together, max), benchmark by benchmark
On GPQA Diamond, Kimi K3.1 Pro (Together, max) scores 94.2% and ranks #12 of 302. Only 11 models score higher, led by GPT-6.2 Astra (max) at 96.8%, which is 2.7 points ahead. It covers expert-level science questions. The board median is 75.1%, so Kimi K3.1 Pro (Together, max) is 19.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
On OTIS Mock AIME, Kimi K3.1 Pro (Together, max) scores 100.0% and ranks #12 of 206. Only 11 models score higher, led by Claude Fable 5.1 (max) at 100.0%, which is 0.0 points ahead. It covers competition maths (AIME 2024-25 style). The board median is 66.4%, so Kimi K3.1 Pro (Together, max) is 33.6 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
Kimi K3.1 Pro (Together, max) manages 62.2% on SimpleQA Verified, good for #13 of 83. Claude Fable 5.2 (max) leads at 94.0%, 31.8 points further up, and 84% of models sit below. The results either side are GPT-5.5 (xhigh) (63.0%) and GPT-6 Sol (max) (60.7%). It covers short factual questions. The board median is 41.4%, so Kimi K3.1 Pro (Together, max) is 20.8 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
FrontierMath T1-3 puts Kimi K3.1 Pro (Together, max) at 78.2%, #21 of 78. That beats 73% of the models tested, but 16.8 points short of Claude Fable 5.2 (max) (95.1%). The results either side are GPT-5.4 (xhigh) (78.6%) and Claude Haiku 5.5 (max) (75.1%). It covers research-level maths, tiers 1-3. The board median is 61.6%, so Kimi K3.1 Pro (Together, max) is 16.7 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
On Mystery Games, Kimi K3.1 Pro (Together, max) reaches 42.0%. It ranks #21 of 69, ahead of 70% of the field, and 53.0 points behind GPT-6.2 Astra (max) at 95.0%. The results either side are deepseek-v4.1-flash-max (43.0%) and Qwen3.8 Max (xhigh) (38.0%). It covers deductive reasoning puzzles. The board median is 26.0%, so Kimi K3.1 Pro (Together, max) is 16.0 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
Kimi K3.1 Pro (Together, max) manages 45.0% on Chess Puzzles, good for #26 of 84. Claude Fable 5.2 (max) leads at 79.0%, 34.0 points further up, and 69% of models sit below. The results either side are GPT-5.4 nano (xhigh) (45.0%) and Claude Opus 5 (max) (43.0%). It covers chess tactics puzzles. The board median is 33.0%, so Kimi K3.1 Pro (Together, max) is 12.0 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
FrontierMath T4 puts Kimi K3.1 Pro (Together, max) at 65.9%, #19 of 60. That beats 68% of the models tested, but 34.1 points short of Claude Fable 5.2 (max) (100.0%). The results either side are Muse Wonder (max) (70.7%) and GPT-5.6 Luna (max) (61.0%). It covers research-level maths, tier 4. The board median is 32.9%, so Kimi K3.1 Pro (Together, max) is 32.9 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
On Furniture Assembly, Kimi K3.1 Pro (Together, max) reaches 55.8%. It ranks #14 of 36, ahead of 61% of the field, and 34.2 points behind Claude Fable 5.2 (max) at 90.0%. The results either side are GPT-5.6 Sol (max) (56.7%) and GPT-5.6 Terra (max) (54.2%). It covers planning a physical assembly. The board median is 44.2%, so Kimi K3.1 Pro (Together, max) is 11.7 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.
Strengths and weaknesses
Reasoning, maths and science: strong, ahead of 83% of the field on average across 4 benchmarks. Its best board here is GPQA Diamond (#12 of 302) and its weakest is FrontierMath T4 (#19 of 60). General knowledge and chat: solid, ahead of 71% of the field on average across 4 benchmarks. Its best board here is SimpleQA Verified (#13 of 83) and its weakest is Furniture Assembly (#14 of 36).
Price and value
The median input price across the 443 priced models is $0.482 per million tokens, so Kimi K3.1 Pro (Together, max) costs less than the typical listing. Models that score within 1.5 points or better and cost less: DeepSeek V4.1 Flash 1017 (max) ($0.036, 55.5), Claude Haiku 5.5 (max) ($0.100, 54.3), Kimi K3 (max) ($0.192, 55.2). Similar scores cost more on Muse Spark Open (xhigh) ($1.25) and Muse Spark 1.3 (max) ($1.25).
Release history
The Moonshot model released before it was Kimi K3.1 Pro (max) on 2026-10-23, which scores 55.5.
Reading these numbers
The 8 scores on this page come from 1 source: Epoch AI Benchmarking Hub (8). No official leaderboard supplied any of them, so treat them as third-party measurements.
Models ranked nearby
- #20 DeepSeek V4.1 Flash 1017 (max) (55.5)
- #21 Kimi K3.1 Pro (max) (55.5)
- #22 gemini-3.1-pro-preview-high (55.3)
- #24 Kimi K3 (max) (55.2)
- #25 Gemini 4 Hydrogen (xhigh) (54.7)
- #26 Claude 4 Sonnet (54.5)
More from Moonshot
- Kimi K3.1 Pro (max) (#21, 55.5)
- Kimi K3 (max) (#24, 55.2)
- Kimi K2.5 (Fireworks) (#68, 52.6)
- Kimi K2.5 (#243, 50.3)
- Kimi K2 Thinking Turbo (#245, 50.3)
- Kimi K2 Thinking (Together) (#278, 49.7)
Compare it on the live leaderboard → · How the aggregate score works