Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | gpt-oss-120b (high) | Claude Opus 4.6 (Non-reasoning, High Effort) |
|---|---|---|
| Input ($/M tokens) | $0.15 | $5 |
| Output ($/M tokens) | $0.6 | $25 |
Verdict. gpt-oss-120b (high) wins the overall benchmark matchup 6–5 across 11 overlapping categories, but raw benchmark score is only one input to the decision.
Pricing. Both models sit in the budget / premium bracket for output-token pricing. At 0.0× the per-million-token cost, gpt-oss-120b (high) is meaningfully cheaper if your traffic is output-heavy (long completions, document generation, agent loops). gpt-oss-120b (high) makes more sense when output volume is low and absolute reasoning quality justifies the premium.
Strengths. gpt-oss-120b (high) is strongest on AIME 2025 (93%), Math Index (93.4), LiveCodeBench (88%). Claude Opus 4.6 (Non-reasoning, High Effort) leads on GPQA Diamond (84%), TerminalBench (48%), SciCode (46%).
Speed. On throughput, gpt-oss-120b (high) generates tokens at 217 tok/s versus 0 tok/s — about 100% faster. On time-to-first-token, Claude Opus 4.6 (Non-reasoning, High Effort) responds in 0ms vs 535ms, which matters most for chat-style UIs.
Provider. OpenAI and Anthropic sell to overlapping but distinct developer audiences: OpenAI tends to ship frontier reasoning models with premium positioning, while Anthropic 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): gpt-oss-120b (high) costs $13.50 ($162/year); Claude Opus 4.6 (Non-reasoning, High Effort) costs $525.00 ($6300/year). At a smaller 5M-input/2M-output scale (single-developer tool or prototype): gpt-oss-120b (high) ≈ $1.95/run, Claude Opus 4.6 (Non-reasoning, High Effort) ≈ $75.00/run. At agent/realtime scale (200M input / 100M output per million requests): gpt-oss-120b (high) ≈ $90/run, Claude Opus 4.6 (Non-reasoning, High Effort) ≈ $3500/run. gpt-oss-120b (high) becomes more attractive at higher volume — the absolute per-token pricing difference compounds when you ship at scale.
Recommendation. Benchmarks are too close to call (1-win swing) but Claude Opus 4.6 (Non-reasoning, High Effort) is 41.7× more expensive per million output tokens. For most workloads the cheaper option (gpt-oss-120b (high)) wins on cost-quality tradeoff. Only pay the premium for Claude Opus 4.6 (Non-reasoning, High Effort) if you have a measured lift on a task your product depends on.
Head-to-head deltas
Data from Artificial Analysis API — 12 benchmarks
gpt-oss-120b (high) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.26/M tokens vs $10.00/M for Claude Opus 4.6 (Non-reasoning, High Effort).
gpt-oss-120b (high) wins 6 out of 12 benchmarks compared to 5 for Claude Opus 4.6 (Non-reasoning, High Effort). See the detailed benchmark chart above for per-category results.
gpt-oss-120b (high) generates tokens faster at 217 tok/s vs 0 tok/s. However, Claude Opus 4.6 (Non-reasoning, High Effort) has lower time-to-first-token (0.00s vs 0.54s).
Choose based on your priorities: gpt-oss-120b (high) for lower cost, gpt-oss-120b (high) for stronger benchmark performance, and gpt-oss-120b (high) for faster generation. For latency-sensitive apps, check the TTFT comparison above.