GPT-5.3: token counter & pricing
OpenAI · exact (uses official tokenizer) · pricing as of 2026-05-31.
- Provider
- OpenAI
- API model ID
gpt-5.3-chat-latest- Context window
- 400,000 tokens
- Input price
- $1.75 per 1M tokens
- Output price
- $14.00 per 1M tokens
- Tokenizer accuracy
- exact (uses official tokenizer)
- Pricing as of
- 2026-05-31
Open the counter to count tokens for GPT-5.3 in real time.
What is GPT-5.3?
GPT-5.3 is the third revision in the GPT-5 line, available in two specialized variants: GPT-5.3 Chat (general conversation, the default gpt-5.3-chat-latest) and GPT-5.3 Codex (coding-tuned, gpt-5.3-codex). Both share the same per-token price: $1.75 input / $14 output per 1M tokens, identical to GPT-5.2 but with material upgrades on conversational and coding benchmarks respectively.
This counter uses the Chat variant as the canonical GPT-5.3 entry since both variants tokenize identically.
How tokens are counted here
OpenAI's o200k_base tokenizer. Browser-side via js-tiktoken. Exact.
Pricing notes
$1.75 input / $14 output per 1M. Cached input $0.175/M.
Same price tier as GPT-5.2, the choice between them is about model behavior, not cost. GPT-5.3's specialized tuning is the differentiator.
When to use GPT-5.3
- Conversational AI workloads (chatbots, support, multi-turn agents). Chat variant is measurably better on dialog coherence than 5.2.
- Coding workloads. Codex variant outperforms 5.2 on most coding benchmarks (HumanEval, SWE-Bench).
- Specialized routing in agent systems, use Chat for conversational nodes, Codex for code-gen nodes, paying the same per-token rate.
When not to use it:
- General workloads with no conversational or coding lean, base GPT-5.2 is functionally equivalent.
- Frontier-tier reasoning. GPT-5.4 or 5.5.
Common questions
Should I use Chat or Codex?
Pick by primary workload. Mixed workloads: route per-task. The pricing is identical, so there's no penalty for using both in the same pipeline.
How does GPT-5.3 Codex compare to Anthropic's coding flagship?
Anthropic's coding sweet spot is Claude Sonnet 4.6 at $3/$15, 40% more expensive on output than 5.3 Codex. Sonnet often wins on reading large existing codebases; 5.3 Codex often wins on greenfield generation. Both are competitive, run your own evals.
Cached input on Chat vs Codex?
Same caching rate ($0.175/M). The cached-prompt discount works identically across both variants.
Compare GPT-5.3 to other models
- GPT-5.5 (OpenAI, $5.00/$30.00)
- GPT-5.5 Pro (OpenAI, $30.00/$180.00)
- GPT-5.4 (OpenAI, $2.50/$15.00)
- GPT-5.4 Mini (OpenAI, $0.75/$4.50)
- GPT-5.4 Nano (OpenAI, $0.20/$1.25)
- GPT-5.4 Pro (OpenAI, $30.00/$180.00)
- GPT-5.2 (OpenAI, $1.75/$14.00)
- GPT-5.2 Pro (OpenAI, $21.00/$168.00)
- GPT-5.1 (OpenAI, $1.25/$10.00)
- GPT-5 (OpenAI, $1.25/$10.00)
- GPT-5 Mini (OpenAI, $0.25/$2.00)
- GPT-5 Nano (OpenAI, $0.05/$0.40)
- GPT-5 Pro (OpenAI, $15.00/$120.00)
- GPT-4.1 (OpenAI, $2.00/$8.00)
- GPT-4.1 Mini (OpenAI, $0.40/$1.60)
- GPT-4.1 Nano (OpenAI, $0.10/$0.40)
- o3 (OpenAI, $2.00/$8.00)
- o3-mini (OpenAI, $1.10/$4.40)
- o3-pro (OpenAI, $20.00/$80.00)
- o4-mini (OpenAI, $1.10/$4.40)
- GPT-4o (OpenAI, $2.50/$10.00)
- GPT-4o mini (OpenAI, $0.15/$0.60)
- GPT-4 Turbo (OpenAI, $10.00/$30.00)
- Gemini 3.1 Pro (Google, $2.00/$12.00)
- Mistral Large (Mistral, $2.00/$6.00)
- Qwen3 Coder 480B (Alibaba, $2.00/$2.00)