Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | Celeris-1 | Qwen3 Next 80B A3B (Reasoning) |
|---|---|---|
| Input ($/M tokens) | $2 | $0.5 |
| Output ($/M tokens) | $6 | $6 |
Verdict. Qwen3 Next 80B A3B (Reasoning) takes the aggregate benchmark matchup 10–0 across 10 categories. Real workloads usually care about a handful of specific tasks — see the per-benchmark table above.
Pricing. Both models sit in the mid-tier bracket for output-token pricing. At 1.0× the per-million-token cost, Qwen3 Next 80B A3B (Reasoning) is meaningfully cheaper if your traffic is output-heavy (long completions, document generation, agent loops). Qwen3 Next 80B A3B (Reasoning) makes more sense when output volume is low and absolute reasoning quality justifies the premium.
Strengths. Celeris-1 is strongest on GPQA Diamond (63%), SciCode (21%), Coding Index (14.4). Qwen3 Next 80B A3B (Reasoning) leads on AIME 2025 (84%), Math Index (84.3), LiveCodeBench (78%).
Speed. On throughput, Celeris-1 generates tokens at 2195 tok/s versus 206 tok/s — about 91% faster. On time-to-first-token, Celeris-1 responds in 482ms vs 1085ms, which matters most for chat-style UIs.
Provider. Celeris and Alibaba sell to overlapping but distinct developer audiences: Celeris 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): Celeris-1 costs $150.00 ($1800/year); Qwen3 Next 80B A3B (Reasoning) costs $105.00 ($1260/year). At a smaller 5M-input/2M-output scale (single-developer tool or prototype): Celeris-1 ≈ $22.00/run, Qwen3 Next 80B A3B (Reasoning) ≈ $14.50/run. At agent/realtime scale (200M input / 100M output per million requests): Celeris-1 ≈ $1000/run, Qwen3 Next 80B A3B (Reasoning) ≈ $700/run. Qwen3 Next 80B A3B (Reasoning) becomes more attractive at higher volume — the absolute per-token pricing difference compounds when you ship at scale.
Recommendation. If you want one safe default, take Qwen3 Next 80B A3B (Reasoning) — it dominates the benchmark table and the latency profile is 10.7× faster. Celeris-1 only makes sense when you specifically need its pricing tier, an existing contract, or a feature difference that is not measured by the benchmarks above.
Head-to-head deltas
Data from Artificial Analysis API — 12 benchmarks
Qwen3 Next 80B A3B (Reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $1.88/M tokens vs $3.00/M for Celeris-1.
Qwen3 Next 80B A3B (Reasoning) wins 10 out of 12 benchmarks compared to 0 for Celeris-1. See the detailed benchmark chart above for per-category results.
Celeris-1 generates tokens faster at 2195 tok/s vs 206 tok/s. Celeris-1 also has lower time-to-first-token (0.48s vs 1.08s).
Choose based on your priorities: Qwen3 Next 80B A3B (Reasoning) for lower cost, Qwen3 Next 80B A3B (Reasoning) for stronger benchmark performance, and Celeris-1 for faster generation. For latency-sensitive apps, check the TTFT comparison above.