The September 2026 price war moved the floor of the LLM API market to $0.10 per million input tokens and $0.50 per million output tokens, and put three vendors’ mid-tier models at the same $2 per million input price. Prices are as of 2026-09-28 (US dollars per million tokens, standard tier, short context), taken from each vendor’s pricing page, with the OpenAI page rechecked on 2026-09-29. Prices in this market move weekly: check the vendor page before committing to a budget. The only secondary source is Simon Willison’s Opus and Sol and Luna post, cited for the price-change history and for older model prices that no longer appear on the vendor pages.
TL;DR
- New floor: OpenAI GPT-6 Luna at $0.10/$0.50 per million tokens is one tenth the price of Anthropic Haiku 4.5 ($1/$5). Simon Willison notes that among OpenAI models only the older, weaker GPT-4.1 Nano ($0.10/$0.40, April 2025) and GPT-5 Nano ($0.05/$0.40, August 2025) were cheaper.
- New mid-tier: GPT-6 Sol (now listed as GPT-6.1 Sol) at $2/$10, Grok 4.7 at $2/$6, and Claude Sonnet 5 and Sonnet 5.5 at $2/$10 now sit at the same dollar-per-million; the difference is output quality, not sticker price.
- Frontier: Claude Opus 5.5 at $4/$20 is a 20 percent cut from Opus 4.x ($5/$25). Claude Fable 5.1 at $10/$50 and GPT-6 Astra at $10/$50 sit on top.
- Cache read cuts: Anthropic dropped the cache-read multiplier from 10 percent to 5 percent on Opus 5.5 (and to 2.5 percent on Fable 5.1). A cache hit now costs one twentieth of the base input price on Opus 5.5 and one fortieth on Fable 5.1, against one tenth before.
- November 2026 cliff: OpenAI’s pricing page says GPT-5.6 Sol’s promotional pricing is available at least through 2026-11-21, and Simon Willison reports a scheduled 25 percent price increase for GPT-5.6 in November. Workloads on GPT-5.6 have until then to move.
- Local now wins on privacy more than on price. At the new floor, the monthly volume at which self-hosted hardware undercuts the API has moved sharply up; see the cost-crossover article and the worked numbers below.
What “after the September price war” actually means
Three price moves landed within two days in September 2026: Grok 4.7 on 21 September, then Claude Opus 5.5 and GPT-6 Sol and Luna on 22 September. Anthropic cut Opus 5.5 to $4 per million input tokens and $20 per million output tokens, a 20 percent drop from the $5/$25 that Opus 4.5, 4.6, 4.7, 4.8, and the original Opus 5 had shared since launch. OpenAI released GPT-6 Sol and GPT-6 Luna alongside the existing GPT-6 Astra (Sol at $2/$10, Luna at $0.10/$0.50, Astra at $10/$50); Simon Willison’s price table puts GPT-6 Sol and Luna at half the price of their GPT-5.6 equivalents.
The most under-reported move is the Anthropic cache-read reduction. A cache hit on Opus 5.5 costs $0.20 per million tokens, which is 5 percent of the base input price rather than the 10 percent that every prior Claude model used. On Fable 5.1 the multiplier drops to 2.5 percent, so a cache hit costs $0.25 per million tokens on a $10 base.
The new floor: GPT-6 Luna versus Haiku 4.5
The cheapest current API model on each vendor’s pricing page is a useful anchor. On 2026-09-28 the anchors are:
| Model | Input $/MTok | Output $/MTok | Cached input $/MTok | Source |
|---|---|---|---|---|
| Claude Haiku 4.5 | $1.00 | $5.00 | $0.10 | Anthropic pricing |
| OpenAI GPT-6 Luna | $0.10 | $0.50 | $0.01 | OpenAI pricing |
Luna is ten times cheaper than Haiku 4.5 on input and on output, and ten times cheaper than Haiku’s own cache read. Simon Willison notes that GPT-4.1 Nano in April 2025 was $0.10/$0.40 and GPT-5 Nano in August 2025 was $0.05/$0.40; GPT-6 Luna at $0.10/$0.50 matches GPT-4.1 Nano on input and raises the output price 25 percent above both prior Nanos. For bulk classification, log triage, embedding-style scoring, and “summarise this transcript” workloads, the new floor is cheap enough that the choice is no longer whether to send the request but how many workers to fan it out across.
The mid-tier sweet spot: GPT-6 Sol, Grok 4.7, Sonnet 5
Three models from three vendors now share the $2 input price point, which makes the mid-tier a quality comparison rather than a price comparison. From the vendor pricing pages on 2026-09-28:
| Model | Input $/MTok | Output $/MTok | Cached input $/MTok | Source |
|---|---|---|---|---|
| Claude Sonnet 5 | $2.00 | $10.00 | $0.20 | Anthropic pricing |
| OpenAI GPT-6.1 Sol | $2.00 | $10.00 | $0.10 | OpenAI pricing |
| xAI Grok 4.7 (under 200k) | $2.00 | $6.00 | $0.50 | xAI pricing |
OpenAI’s pricing page listed GPT-6.1 Sol in place of GPT-6 Sol when rechecked on 2026-09-29; Simon Willison’s 22 September table had GPT-6 Sol at the same $2/$10 with $0.20 cached input. Grok 4.7 is the cheapest on output at $6 per million tokens; the others are at $10. The reverse is true on cache reads: GPT-6.1 Sol is at 5 percent of input ($0.10), Sonnet 5 at 10 percent ($0.20), and Grok 4.7 at 25 percent ($0.50). For cache-heavy retrieval, GPT-6.1 Sol and Sonnet 5 are the cheaper choice, while Grok 4.7 wins on raw output for short, non-cached calls.
Grok 4.7 also has a two-tier context price: $4 input, $12 output, and $1.00 cached for requests of 200k tokens and above. OpenAI also publishes separate long-context rates for its GPT-6 models (GPT-6.1 Sol $4/$15 long context).
The frontier tier: Opus 5.5 versus Fable 5.1 versus GPT-6 Astra
Three models currently sit at the top of each vendor’s catalogue. From the vendor pricing pages on 2026-09-28:
| Model | Input $/MTok | Output $/MTok | Cached input $/MTok | Source |
|---|---|---|---|---|
| Claude Opus 5.5 | $4.00 | $20.00 | $0.20 | Anthropic pricing |
| Claude Fable 5.1 | $10.00 | $50.00 | $0.25 | Anthropic pricing |
| OpenAI GPT-6 Astra | $10.00 | $50.00 | $1.00 | OpenAI pricing |
Opus 5.5 at $4/$20 is the cheap frontier. Fable 5.1 and GPT-6 Astra are the same price on input and output ($10/$50) but Fable 5.1 has a 2.5 percent cache read multiplier ($0.25), while GPT-6 Astra is at 10 percent ($1.00). Anthropic also publishes a fast-mode premium for Opus 5.5 of $8/$40 per million, doubling the input and output rates for lower-latency inference.
The capability story is not in this article; the capability ceiling piece covers benchmarks, licence terms, and context windows for the same model set, refreshed on 2026-09-27.
Cached input is the real discount
The price war is more interesting in cache than in sticker. From the Anthropic pricing page:
- Opus 5.5: cache hit at 5 percent of base ($0.20 per million, against a $4 base).
- Fable 5.1: cache hit at 2.5 percent of base ($0.25 per million, against a $10 base).
- All other Anthropic models: cache hit at 10 percent of base ($1 per million on Fable 5, $0.50 per million on Opus 5, and so on).
Simon Willison’s post calls the 60 percent fall in Opus cache-read pricing significant for longer agentic conversations (Opus 5 was $0.50 per million; Opus 5.5 is $0.20). For a 50,000-token system prompt with a 1-hour cache on Opus 5.5, the prompt costs about $0.40 to write (50,000 tokens at $8/MTok for the 2x 1-hour write multiplier) and $0.01 to read on every subsequent request (50,000 tokens at $0.20/MTok). At a thousand re-reads the prompt amortises to about one cent per call.
On the OpenAI side, GPT-6 Luna and GPT-6 Astra price cache reads at 10 percent of base, and GPT-6.1 Sol at 5 percent (GPT-5.6 Sol is at 10 percent). Most of the OpenAI gain is the underlying base-price drop.
Where local inference still beats the API
The cost crossover has shifted since August. Per the cost-crossover article, the monthly token volume at which a self-hosted model becomes cheaper than the cheapest API option depends on three numbers: the amortised hardware cost, the electricity draw, and the inference throughput. After the September cuts the API floor dropped from $1/$5 (Haiku 4.5) to $0.10/$0.50 (GPT-6 Luna), so the crossover point moved up.
Two cases are worth separating:
-
High-volume workloads. As an illustration, a card generating 30 tokens per second around the clock (an assumed rate, not a measurement) produces about 78 million tokens a month. The same 78 million output tokens cost about $39 at the GPT-6 Luna output price of $0.50 per million, which is in the same range as the electricity for a GPU running 24/7. At the new floor, volume alone rarely justifies local hardware against the cheapest API tier; the price case for local depends on hardware already owned and on which API tier the workload actually needs.
-
Workloads that cannot leave the machine. Per the privacy hub, prompts that contain code, customer data, or unreleased product information should not flow through a hosted endpoint by default. Local inference with a Qwen3 or Llama 3.3 open-weight model keeps the bytes on the box.
For workloads below the crossover that are safe to send to the API, GPT-6 Luna is the cheapest option on current list prices. For workloads above the crossover or that must stay local, the best-local-LLM guide and the cost-crossover piece cover the hardware ladder in detail.
The November 2026 cliff
Simon Willison flagged on 2026-09-22 that GPT-5.6 has a scheduled 25 percent price increase for November. The OpenAI pricing page says GPT-5.6 Sol’s promotional pricing is available at least through 2026-11-21 and does not state the later rate. Simon Willison’s GPT-5.6 Sol row lists $4/$20, which is exactly the new Opus 5.5 rate; a 25 percent rise would put GPT-5.6 Sol at $5/$25, the price Opus 4.5 through Opus 5 carried.
Two operational notes:
- Migration window: any production workload on GPT-5.6 has roughly eight weeks to either accept the 25 percent rise or migrate to GPT-6.1 Sol ($2/$10 on 2026-09-29).
- Date risk: OpenAI says “at least through” 2026-11-21, so the promotion may run longer. The $5/$25 figure is arithmetic on the reported 25 percent rise, not a published price; verify on the vendor pricing page before committing.
The Bottom Line
After the September 2026 cuts the API market has three usable tiers and one sub-tier: cheap (GPT-6 Luna at $0.10/$0.50), mid (Sonnet 5 and 5.5, GPT-6.1 Sol, Grok 4.7 at $2 with output from $6 to $10), and frontier (Opus 5.5 at $4/$20, Fable 5.1 and GPT-6 Astra at $10/$50). The cache-read multiplier is where the real saving lives on long-system-prompt workloads, and the November 2026 GPT-5.6 price rise is the date to plan around. Local inference still wins on workloads that must stay on the machine; for everything else, GPT-6 Luna is the cheapest list price and the mid-tier is a quality choice at near-identical prices. All prices as of 2026-09-28; check the vendor pages before relying on them.
Related on Intelligibberish
- Leading Frontier AI Models: Capability Ceiling (September 2026) - the capability ceiling, refreshed 2026-09-27, with the same models at the same vendor pricing.
- When Local AI Beats the API: The Cost Crossover (August 2026) - the monthly volume at which a self-hosted model crosses below the API, with the hardware and electricity assumptions.
- Best Local LLM for Everyday Tasks - the open-weight ladder from 8 GB laptops to 32 GB workstations.