readable AI benchmarks (simplified)
Kimi
Kimi K3 (Max)
Kimi K3 (Max) has a Quality Score of 63.5, ranking 45th among 572 scored models. Per 1M tokens, pricing is $3.00 input, $15.00 output, and $0.300 cached input. It accepts text, image, and video input, outputs text, and has a 1,048,576 token context window. Its Reliability Score is 59.9, ranking 67th among 574 scored models, above average (total average: 36.5). Its Value Score is 40.2, high compared with the total Value average of 31.5.
Model specification
- Reasoning
- reasoning
- Input modalities
- text, image, video
- Output modalities
- text
- Context window
- 1,048,576 tokens
- Weights
- open weights
Token prices USD per 1M tokens
- Input
- $3.00
- Output
- $15.00
- Cached Input
- $0.300
- Cache write
- not available
- Cost per task
- $2.00
Capability
- Intelligence
- 43.6
- Coding
- not available
- Agentic
- not available
- Omniscience
- 19.7
- Correct
- 47.6%
- Blended price
- $2.31
Answer outcomes
Fully graded outcomes (Correct + Incorrect): 75.5%
Artificial Analysis benchmarks
- GDPval-AA v2
- 52%
- τ³-Banking
- 46%
- SciCode
- 59%
- Humanity’s Last Exam
- 47%
- GPQA Diamond
- 94%
- CritPt
- 23%
- AA-Omniscience
- 60%
- AA-LCR
- 89%
DeepSWE operational cross-check
- Pass@1
- 68.5%
- Avg steps/scored attempt
- 97.59
- Mean cost/scored attempt (USD)
- $4.65
- Mean output tokens/scored attempt
- 81,499.843
- Mean duration/scored attempt (s)
- 4541 s
Mean cost/scored attempt is the official feed's mean_cost_usd: mean benchmark cost in USD per scored attempt under the source's cost basis. It is not Artificial Analysis cost per task and not token price (USD per 1M tokens). Mean output tokens/scored attempt is mean_output_tokens, also averaged over every scored attempt including scored failures. Mean duration/scored attempt is mean_duration_seconds; it is informational only, and wall-clock varies with provider and host load. These three fields do not affect Quality Score, Value Score, or Reliability. Epoch General ECI's benchmark basket includes DeepSWE v1.1, so Epoch General ECI and this DeepSWE cross-check are not independent evidence.
Similar models
- Kimi K3 (Low)KimiQuality Score 49.1Value Score 31.1
- Kimi K2.6 (Reasoning)KimiQuality Score 47.1Value Score 34.4
- Kimi K2.7 CodeKimiQuality Score 42.3Value Score 30.8
- Kimi K2.5 (Reasoning)KimiQuality Score 41.2Value Score 30.4
- Kimi K2 ThinkingKimiQuality Score 36.6Value Score 26.3
- Kimi K2.6 (Non-reasoning)KimiQuality Score 40.8Value Score 29.8