Cost mode:

Category: Structured Data & Fact Extraction · Rail: absolute · Typical I/O: 2334→9656 tokens

Models

Frontier on this task: Qwen 3.6 Plus at 9.84 / 10. Quality bar at 95%: 9.35.

024681095% barQwen 3.5 Flash$0.002581/call87% cheaperDeepSeek V4 Flash$0.003030/call85% cheaperMiniMax M2.5$0.012287/call37% cheaperGPT-5.4 nano$0.012537/call36% cheaperQwen 3.6 Plus$0.019588/call0% cheaperDeepSeek V4 Pro$0.037664/call-92% cheaperKimi K2.6$0.040841/call-109% cheaperGemini 3.1 Pro Preview$0.120540/call-515% cheaperGPT-5.5$0.301350/call-1438% cheaperHaiku 4.5$0.050614/call-158% cheaperClaude Opus 4.7$0.253070/call-1192% cheaperClaude Sonnet 4.6$0.151842/call-675% cheaperGemini 3 Flash Preview$0.030135/call-54% cheaperGemini 3.1 Flash Lite$0.015068/call23% cheaperGPT-5.4 mini$0.045202/call-131% cheaper

point-estimate floor (CI low) · upper CI (less certain) · Bars sorted by blended cost; cheapest qualifier first. Greyed rows are MEDIUM+ models whose point estimate clears the bar but whose CI low does not.

Cost breakdown

ModelQualitySampleBlended cost / callSavings vs bestMode
Qwen 3.5 Flash Alibaba Cloud (DashScope)9.78 / 10 CI [9.69, 9.87]n=100 · high$0.00258187% cheapersync
DeepSeek V4 Flash DeepSeek9.75 / 10 CI [9.65, 9.85]n=88 · ranked$0.00303085% cheapersync
MiniMax M2.5 MiniMax9.71 / 10 CI [9.34, 10.00]n=85 · medium$0.01228737% cheapersync
GPT-5.4 nano OpenAI9.76 / 10 CI [9.54, 9.99]n=88 · high$0.01253736% cheaperbatch
Qwen 3.6 Plus best Alibaba Cloud (DashScope)9.84 / 10 CI [9.69, 9.99]n=100 · ranked$0.019588(anchor)sync
DeepSeek V4 Pro DeepSeek9.43 / 10 CI [9.22, 9.64]n=89 · high$0.037664sync
Kimi K2.6 Moonshot AI9.54 / 10 CI [9.32, 9.74]n=99 · high$0.040841batch
Gemini 3.1 Pro Preview Gemini9.63 / 10 CI [9.41, 9.85]n=90 · high$0.120540batch
GPT-5.5 OpenAI9.42 / 10 CI [9.15, 9.70]n=81 · high$0.301350batch

Typical call shape for this task: 2334 input tokens → 9656 output tokens, EMA-tracked from production traffic. Blended cost = (in × in_price + out × out_price), rounded to 6 decimals.