- Rank
- #334
- Total Output Tokens
- 57,866
- Response Time (avg)
- 22.44s
- Total Cost
- $0.119
DeepSeek V3.2 vs KAT Coder AIR V2.5 (default)
DeepSeek V3.2 leads on average score with 4.6 vs 4.4. KAT Coder AIR V2.5 (default) has the lower benchmark cost at $0.070 vs $0.119. KAT Coder AIR V2.5 (default) is faster at 11.96s vs 22.44s, with pass rates of 34.8% vs 34.8%.
Compared models
- Rank
- #348
- Total Output Tokens
- 97,752
- Response Time (avg)
- 11.96s
- Total Cost
- $0.070
Recommended model
KAT Coder AIR V2.5 (default)
Its score stays close to the best score here (4.4 vs 4.6), while costing about 1.7x less than DeepSeek V3.2.
Detailed comparison
| Metric | DeepSeek V3.2 DeepSeek V3.2 none | KAT Coder AIR V2.5 KAT Coder AIR V2.5 default |
|---|---|---|
| Score | 4.6 | 4.4 |
| Rank | #334 | #348 |
| Reliability | 10.0 | 10.0 |
| Consistency | 8.2 | 7.4 |
| Attempts | 69/69 | 69/69 |
| Tests Correct | ||
| Attempt pass rate | 34.8% | 34.8% |
| Flaky tests | 5 | 7 |
| Total Runs | 69 | 69 |
| Cost per result | 1.943 | 1.382 |
| Total Cost | $0.119 | $0.070 |
| Input Price | $0.280 / 1M | $0.150 / 1M |
| Output Price | $0.420 / 1M | $0.600 / 1M |
| Cache Read Price | $0.028 / 1M | N/A |
| Cache Write Price | N/A | N/A |
| Total Input Tokens | 342,158 | 69,376 |
| Output Tokens | 32,853 | 7,792 |
| Reasoning Tokens | 25,013 | 89,960 |
| Response Time (avg) | 22.44s | 11.96s |
| Response Time (max) | 122.50s | 121.05s |
| Response Time (total) | 516.03s | 275.18s |
| Parameters | 671B total (37B active) | ~35B total (~3B active) |
| Availability | Open source | Closed |
Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.
Model generation showcase
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
#334 DeepSeek V3.2
none- Cost
- $0.002
- Time
- 7.0s
- Tokens
- 1,046 tok
#348 KAT Coder AIR V2.5
default- Cost
- $0.002
- Time
- 14.5s
- Tokens
- 2,619 tok
Top Models by Score
Score vs Total Cost
Response Time (avg)
Score vs Response Time (avg)
Total Output Tokens
Score vs Total Output Tokens
Category Breakdown
| Agentic | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 3.0 | 10.0 | 0.0% | 0 | 122.50s | 206,327 | 15,767 | 0 | |
| KAT Coder AIR V2.5 | 3.0 | 10.0 | 0.0% | 0 | 1.07s | 0 | 0 | 0 |
| Anti-AI Tricks | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 3.2 | 8.0 | 8.3% | 1 | 9.35s | 494 | 1,073 | 0 | |
| KAT Coder AIR V2.5 | 5.3 | 5.6 | 50.0% | 2 | 2.68s | 672 | 686 | 2,358 |
| Coding | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 3.1 | 6.9 | 11.1% | 1 | 14.54s | 7,279 | 4,528 | 0 | |
| KAT Coder AIR V2.5 | 3.3 | 9.6 | 0.0% | 0 | 14.31s | 6,472 | 2,032 | 14,886 |
| Combined | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 4.8 | 10.0 | 0.0% | 0 | 113.53s | 109,626 | 10,550 | 23,269 | |
| KAT Coder AIR V2.5 | 3.8 | 1.6 | 50.0% | 2 | 73.00s | 43,170 | 2,531 | 48,237 |
| Data parsing and extraction | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 6.3 | 5.8 | 66.7% | 1 | 9.42s | 7,890 | 210 | 1,500 | |
| KAT Coder AIR V2.5 | 6.3 | 5.8 | 66.7% | 1 | 2.66s | 7,776 | 280 | 988 |
| Domain specific | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 3.0 | 10.0 | 0.0% | 0 | 1.48s | 675 | 23 | 0 | |
| KAT Coder AIR V2.5 | 3.0 | 10.0 | 0.0% | 0 | 8.33s | 780 | 1,399 | 10,258 |
| General Intelligence | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 4.7 | 1.6 | 66.7% | 1 | 9.32s | 314 | 43 | 0 | |
| KAT Coder AIR V2.5 | 5.0 | 10.0 | 0.0% | 0 | 12.02s | 516 | 141 | 515 |
| Instructions following | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 10.0 | 10.0 | 100.0% | 0 | 1.52s | 627 | 66 | 0 | |
| KAT Coder AIR V2.5 | 9.9 | 10.0 | 100.0% | 0 | 1.75s | 699 | 90 | 768 |
| Puzzle Solving | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 7.6 | 7.2 | 77.8% | 1 | 6.91s | 424 | 298 | 0 | |
| KAT Coder AIR V2.5 | 2.9 | 4.4 | 22.2% | 2 | 1.84s | 696 | 284 | 971 |
| Tool Calling | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 10.0 | 10.0 | 100.0% | 0 | 11.85s | 8,319 | 278 | 244 | |
| KAT Coder AIR V2.5 | 10.0 | 10.0 | 100.0% | 0 | 5.13s | 8,391 | 330 | 1,230 |
| Trivia | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 3.0 | 10.0 | 0.0% | 0 | 17.23s | 183 | 17 | 0 | |
| KAT Coder AIR V2.5 | 3.0 | 10.0 | 0.0% | 0 | 17.96s | 204 | 19 | 9,749 |
Quick Compare
Switch Comparison Pair
KAT Coder AIR V2.5defaultvsNemotron 3 SupernoneFree AvailableGranite 4.2 8BnonevsKAT Coder AIR V2.5defaultDeepSeek V3.2nonevsQwen3 Coder NextmediumDeepSeek V3.2nonevsLaguna S 2.1lowFree AvailableDeepSeek V3.2nonevsGranite 4.2 8BhighDeepSeek V3.2nonevsNemotron 3.5 LightninglowFree AvailableCobuddymediumFree AvailablevsKAT Coder AIR V2.5defaultNorth Mini CodedefaultFree AvailablevsDeepSeek V3.2noneMercury 2nonevsKAT Coder AIR V2.5defaultDeepSeek V3.2nonevsNemotron 3.5 LightningmediumFree AvailableDeepSeek V3.2nonevsMercury 2.5 PreviewdefaultDeepSeek V3.2nonevsLing 2.6 Flashdefault