19 models in this directory cost $1 or less per million input tokens. Gemini 3.6 Flash is the most capable of them ($0.75/1M), Mistral Small 3.1 is the absolute cheapest at $0.1/1M, and GPT-5.6 Luna gives you the largest context window (1.05M tokens) at this price level.
The safest this comparison default, the cheaper option worth trying first, and the specialist pick — before you read the detail below.
Gemini 3.6 Flash is the strongest answer here for this comparison — pick it when quality of output matters more than the $0.75/1M/1M input you pay for it.
DeepSeek V4-Pro handles the same job for about 71% less per token. Start here and only move up if the output is not good enough.
Qwen 3.8 Flash is the fastest of these for this comparison — worth it when latency is what the reader notices, not the last few points of reasoning depth.
Gemini 3.6 Flash is the most capable model under $1/1M — $0.75/1M input, $3.75/1M output, 1.048576M context.
Mistral Small 3.1 is the absolute cheapest at $0.1/1M input — 100x cheaper than Claude Fable 5.1.
DeepSeek V4-Pro is the strongest budget coding pick (coding score 93/100).
Choose Gemini 3.6 Flash as your budget default — the best capability-per-dollar in this price band.
Choose Mistral Small 3.1 for very high-volume tasks like classification, tagging, and extraction.
Route hard tasks to a premium model and keep everything else here — a two-tier setup usually cuts spend 60–80%.
Switch the scoring lens to see whether the this comparison answer changes when cost, speed, or long-document depth leads the decision.
Google / Balanced / Sep 3, 2026
Best Gemini for agents — efficiency king with native computer use.
Ranks models by the broadest mix of coding, writing, research, and long-context usefulness.
You need raw frontier reasoning ceiling — Claude Opus 5 and GPT-5.6 Sol lead the hardest tasks.
Where the this comparison recommendation shifts once you weigh price or latency differently.
Best Gemini for agents — efficiency king with native computer use.
Best open-weights flagship — near-frontier coding at a tenth of the price.
Best budget model from a frontier lab — near-frontier scores at commodity price.
80.8% SWE-bench Verified at introductory Flash pricing.
GPT-4o-class coding quality at under $0.30/1M — the best value in the directory.
Best agentic capability per dollar in the directory.
SWE-bench Pro 62.5 at sixteen cents per million input.
Native vision and video, MIT weights, fifteen cents per million.
List prices and published scores — the numbers this page's pick is built from.
| Model | Input | Output | Est. month | Context | Speed | Coding | Writing | Research |
|---|---|---|---|---|---|---|---|---|
| Gemini 3.6 FlashGoogle | $0.75/1M | $3.75/1M | $15 | 1.0M tokens | Fast | 92 | 86 | 91 |
| DeepSeek V4-ProDeepSeek | $0.43/1M | $0.87/1M | $6.09 | 1M tokens | Balanced | 93 | 80 | 85 |
| GPT-5.6 LunaOpenAI | $0.20/1M | $1.20/1M | $4.40 | 1.1M tokens | Fast | 88 | 85 | 84 |
| Gemini 3.7 FlashGoogle | $0.75/1M | $3.75/1M | $15 | 1.0M tokens | Fast | 89 | 82 | 85 |
Scores out of 100 — how we evaluate models. “Est. month” is 10M in / 2M out at list price: a ceiling, no discounts.
Why each one is on the shortlist for this comparison, what it is genuinely good at, and where we would steer you away from it.
The default answer for this comparison — 92/100 on the coding axis, and the model we would start with unless the price below rules it out.
Google's efficiency-focused successor to 3.5 Flash — higher scores on every benchmark Google tested, ~17% fewer output tokens, and cheaper output pricing.
You need raw frontier reasoning ceiling — Claude Opus 5 and GPT-5.6 Sol lead the hardest tasks.
The best Google model for agents right now. Cheaper, faster, and stronger than 3.5 Flash with the best OSWorld computer-use score in its class. The default Gemini pick until Gemini 4 lands.
Full pricing, benchmark table and release notes on the Gemini 3.6 Flash page.
The value option for this comparison: about 71% less per token than Gemini 3.6 Flash, at 93/100 on coding. Worth starting here and moving up only if the output disappoints.
DeepSeek's 1.6T-parameter (49B active) MoE flagship with hybrid sparse attention — near-frontier coding and reasoning at roughly a tenth of closed-rival pricing, MIT-licensed open weights.
You need vision input, verified agentic performance, or predictable pricing (surge pricing and an announced increase loom).
The open-weights frontier flagship of 2026. Self-reported numbers flatter it and independent agentic scores land lower, but even discounted it's the most capability per dollar in the directory's upper tier — with MIT-licensed weights.
Full pricing, benchmark table and release notes on the DeepSeek V4-Pro page.
Also worth a look for this comparison, at 88/100 on the coding axis.
Best budget model from a frontier lab — near-frontier scores at commodity price. Full GPT-5.6 Luna review →
Rounds out the shortlist for this comparison at 89/100 on coding.
80.8% SWE-bench Verified at introductory Flash pricing. Full Gemini 3.7 Flash review →
UseRightAI recommendations are based on practical decision factors people actually feel in day-to-day use.
Newsletter
We email when the this comparison pick changes, when one of these models moves on price, or when something new displaces the current leader.
No spam. Useful updates only. Affiliate disclosures always clearly labeled.
Gemini 3.6 Flash — $0.75/1M input tokens with the highest capability average in this price band. Best Gemini for agents — efficiency king with native computer use.
Mistral Small 3.1 at $0.1/1M input and $0.3/1M output. It handles ultra-high-volume classification, summarisation, and lightweight vision tasks well despite the price.
DeepSeek V4-Pro, with a coding score of 93/100 at $0.435/1M input.
Budget models trail flagships on hard reasoning, nuanced writing, and complex multi-step coding. Claude Fable 5.1, the current capability leader, scores 99/100 on average vs 90/100 for Gemini 3.6 Flash — use cheap models for volume, not for your hardest work.
Yes — usually 3–5x more. Gemini 3.6 Flash charges $0.75/1M input but $3.75/1M output, so long responses drive the real bill. Our API cost calculator models both sides.