Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | Kimi K2.7 Code | Qwen3 VL 235B A22B (Reasoning) |
|---|---|---|
| Input ($/M tokens) | $0.95 | $0.4 |
| Output ($/M tokens) | $4 | $4 |
Verdict. Kimi K2.7 Code wins the overall benchmark matchup 2–0 across 2 overlapping categories, but raw benchmark score is only one input to the decision.
Pricing. Both models sit in the mid-tier bracket for output-token pricing. At 1.0× the per-million-token cost, Qwen3 VL 235B A22B (Reasoning) is meaningfully cheaper if your traffic is output-heavy (long completions, document generation, agent loops). Qwen3 VL 235B A22B (Reasoning) makes more sense when output volume is low and absolute reasoning quality justifies the premium.
Strengths. Kimi K2.7 Code is strongest on Coding Index (60.8), Intelligence Index (43.0). Qwen3 VL 235B A22B (Reasoning) leads on Intelligence Index (20.9).
Speed. On throughput, Qwen3 VL 235B A22B (Reasoning) generates tokens at 54 tok/s versus 39 tok/s — about 28% faster. On time-to-first-token, Kimi K2.7 Code responds in 2970ms vs 3160ms, which matters most for chat-style UIs.
Provider. Kimi and Alibaba sell to overlapping but distinct developer audiences: Kimi tends to ship frontier reasoning models with premium positioning, while Alibaba often prices more aggressively. Your existing vendor relationships, billing, and SLA preferences may matter as much as the raw numbers above.
Workload cost. Workload scenarios (per million requests at 30M input + 15M output tokens): Kimi K2.7 Code costs $88.50 ($1062/year); Qwen3 VL 235B A22B (Reasoning) costs $72.00 ($864/year). At a smaller 5M-input/2M-output scale (single-developer tool or prototype): Kimi K2.7 Code ≈ $12.75/run, Qwen3 VL 235B A22B (Reasoning) ≈ $10.00/run. At agent/realtime scale (200M input / 100M output per million requests): Kimi K2.7 Code ≈ $590/run, Qwen3 VL 235B A22B (Reasoning) ≈ $480/run. Qwen3 VL 235B A22B (Reasoning) becomes more attractive at higher volume — the absolute per-token pricing difference compounds when you ship at scale.
Recommendation. Both models have legitimate use cases — the right answer depends on whether you are optimizing for benchmark ceiling, latency, or unit cost. Start with the cheaper / faster model, evaluate against your specific task, and only switch if the upgrade shows a meaningful lift.
Data from Artificial Analysis API — 12 benchmarks
Qwen3 VL 235B A22B (Reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $1.30/M tokens vs $1.71/M for Kimi K2.7 Code.
Kimi K2.7 Code wins 2 out of 12 benchmarks compared to 0 for Qwen3 VL 235B A22B (Reasoning). See the detailed benchmark chart above for per-category results.
Qwen3 VL 235B A22B (Reasoning) generates tokens faster at 54 tok/s vs 39 tok/s. Kimi K2.7 Code also has lower time-to-first-token (2.97s vs 3.16s).
Choose based on your priorities: Qwen3 VL 235B A22B (Reasoning) for lower cost, Kimi K2.7 Code for stronger benchmark performance, and Qwen3 VL 235B A22B (Reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.