Cost mode:

Category: Financial Analysis & Trading Decisions · Rail: absolute · Typical I/O: 4242→4314 tokens

Models

Frontier on this task: GPT-5.5 at 8.47 / 10. Quality bar at 95%: 8.05.

024681095% barGPT-5.5$0.075315/call0% cheaperQwen 3.5 Flash$0.001249/call98% cheaperQwen 3.6 Plus$0.009791/call87% cheaperHaiku 4.5$0.012906/call83% cheaperClaude Opus 4.7$0.064530/call14% cheaperClaude Sonnet 4.6$0.038718/call49% cheaperDeepSeek V4 Flash$0.001802/call98% cheaperDeepSeek V4 Pro$0.022394/call70% cheaperGemini 3 Flash Preview$0.007532/call90% cheaperGemini 3.1 Flash Lite$0.003766/call95% cheaperGemini 3.1 Pro Preview$0.030126/call60% cheaperMiniMax M2.5$0.006449/call91% cheaperKimi K2.6$0.012772/call83% cheaperGPT-5.4 mini$0.011297/call85% cheaperGPT-5.4 nano$0.003120/call96% 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
GPT-5.5 best OpenAI8.47 / 10 CI [8.30, 8.65]n=100 · ranked$0.075315(anchor)batch

Typical call shape for this task: 4242 input tokens → 4314 output tokens, EMA-tracked from production traffic. Blended cost = (in × in_price + out × out_price), rounded to 6 decimals.